Chapter 05: Architecture Components (BRD v20.0)
5.0 About This Chapter
This chapter contains the complete Business Requirements Document (BRD) v20.0 for the POS Platform. It serves as the authoritative source for all functional requirements, business rules, state machines, and module specifications.
Authority Rule: When any other chapter in this Blueprint conflicts with content in this chapter, this chapter (BRD) wins.
Module Overview
| Module | Scope | BRD Sections |
|---|---|---|
| 1. Sales | POS workflows, payments, returns, offline operations | 1.1-1.20 |
| 2. Customers | Profiles, groups, tiers, communication | 2.1-2.8 |
| 3. Catalog | Products, variants, pricing, multi-channel | 3.1-3.15 |
| 4. Inventory | Stock management, transfers, counting, RFID | 4.1-4.19 |
| 5. Setup & Configuration | Tenant config, roles, registers, RFID config | 5.1-5.21 |
| 6. Integrations | Shopify, Amazon, Google, payments, shipping | 6.1-6.13 |
| 7. State Machines | All 16+ state machine definitions | 7.1-7.16 |
How to Use This Chapter
- For implementation: Find the relevant Module and Section number
- For business rules: Check the YAML business rules in each module
- For state machines: See Module 7 (State Machine Reference)
- For decisions: 113 decisions documented throughout, summarized in Appendix
- For user stories: Gherkin acceptance criteria at the end of each module
This document consolidates all features into a modular architecture.
- Module 1 (Sales): Covers Scanner, Financials, Discounts, Post-Sale corrections, Gift Cards, Exchanges, Special Orders, Multi-Store, Commissions, Cash Management, Offline Operations, Tax Engine, Payment Integration, and more.
- Module 2 (Customers): Covers Profiles, Merging, Tax Logic, Data Management, Customer Groups, Notes, Communication Preferences, and Privacy Compliance.
- Module 3 (Catalog): Covers Product Types & Data Model (with retail attributes, custom fields, UoM, shipping, templates, matrix management), Product Lifecycle, Pricing Engine (price hierarchy, price books, promotions, markdowns), Barcode Management, Categories/Seasons/Collections, Multi-Channel Management, Shopify Integration, Vendor Management, Search & Discovery, Label Printing, Media Management, Notes & Attachments, Permissions & Approvals, Product Analytics, and comprehensive User Stories with Gherkin acceptance criteria.
- Module 4 (Inventory): Covers Inventory Status Model, Purchase Orders & Procurement, Receiving & Inspection, Reorder Management, Inventory Counting & Auditing, Inventory Adjustments, Inter-Store Transfers, Vendor RMA & Returns, Serial & Lot Tracking, Landed Cost & Costing, Product Movement History, POS & Sales Integration, Online Order Fulfillment, Offline Operations, Alerts & Notifications, Dashboard & Reports, Business Rules, and comprehensive User Stories with Gherkin acceptance criteria.
- Module 5 (Setup & Configuration): Covers System Settings, Multi-Currency, Location Management, Supplier Configuration, User Profiles & Permissions, Time Tracking (Clock-In/Clock-Out), Register Management, Printers & Peripherals, Tax Configuration, UoM Management, Payment Methods, Custom Fields, Approval Workflows, Receipt Configuration, Email Templates, Loyalty Settings, Audit Logging, Consolidated Business Rules (YAML), Tenant Onboarding Wizard, Field Specifications Reference with Technical User Stories, and comprehensive User Stories with Gherkin acceptance criteria.
- Module 6 (Integrations & External Systems): Covers Integration Architecture (provider abstraction, retry/backoff, circuit breaker, idempotency, rate limiting, webhook pipeline), Shopify Integration (enhanced with GraphQL, bulk operations, third-party POS rules, BOPIS, omnichannel), Amazon SP-API Integration (OAuth/LWA, catalog/listings/orders/FBA APIs, FBA+FBM fulfillment, compliance, safety buffers), Google Merchant API Integration (product data management, local inventory, disapproval prevention, image requirements, GBP integration), Cross-Platform Product Data Requirements (unified validation matrix, strictest-rule-wins), Cross-Platform Inventory Sync Rules (safety buffers, oversell prevention, channel-specific rules), Payment Processor Integration, Email Provider Integration, Carrier & Shipping Integration, Integration Hub, Integration Business Rules (YAML), and comprehensive User Stories with Gherkin acceptance criteria.
1. Sales Module
1.1 Core Sales Workflow (Item Entry & Logic)
Scope: Scanner/Manual Entry, Inventory Checks, Parking, and Customer Association.
Cross-Reference: See Module 4, Section 4.13 for inventory reservation model during sales.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant SC as Scanner
participant API as Backend
participant DB as DB
participant PROMO as Promo Engine
Note over U, PROMO: Phase 1: Initiation & Entry
loop Action Loop
U->>UI: Toggle Mode (Sale / Return / Exchange)
alt Input Methods
U->>UI: Scan Barcode / Search SKU
UI->>API: GET /product/{sku}
else Scanner Broadcast
SC->>UI: Tags Detected (Array)
UI->>API: POST /products/bulk-lookup
Note right of API: Max 50 tags per request
end
API->>DB: Fetch Price, Stock, Tax
DB-->>UI: Return Product Data
alt Stock Validation
opt Mode == Sale
UI->>UI: Check Stock > 0
alt Low Stock
UI-->>U: Warning / Block Item
end
end
end
alt Mode Check
opt Mode == Exchange
U->>UI: Load Original Sale for Exchange
UI->>API: GET /orders/{id}
API-->>UI: Return Original Sale Items
UI->>UI: Select Items to Exchange OUT
UI->>UI: Scan/Add New Items IN
UI->>UI: Calculate Price Difference
Note right of UI: Links to Exchange flow in Section 1.4
end
end
UI->>UI: Add to Cart
par Intelligence & Context
UI->>PROMO: Analyze Cart Context
PROMO-->>UI: Trigger Upsell Alert ("Buy 1 more for 10% off")
and Customer Attachment
opt Attach Customer
UI->>API: GET /customer/{id} (Loyalty/Debt/Price Tier)
API-->>UI: Return Profile (Credit Limit, Tax Class, Price Level)
UI->>UI: Recalculate Prices (if Price Tier applies)
UI->>UI: Recalculate Tax (if Exempt)
end
end
end
opt Session Management
U->>UI: Click "Park Sale"
UI->>API: POST /sales/park
API->>DB: Serialize State & Release Locks
U->>UI: Click "Retrieve Sale"
UI->>API: GET /sales/parked
API-->>UI: Restore Cart
end
1.1.1 Parked Sale State Machine
stateDiagram-v2
[*] --> ACTIVE: Cart Created
ACTIVE --> PARKED: Staff Parks Sale
PARKED --> ACTIVE: Staff Retrieves Sale
PARKED --> EXPIRED: TTL Exceeded (4 hours)
EXPIRED --> [*]: Inventory Released
ACTIVE --> PENDING: Proceed to Payment
Parked Sale Rules:
- Maximum parked sales per terminal: 5
- TTL (Time-to-Live): 4 hours
- Inventory is soft-reserved while parked (visible to other terminals with warning)
- Expired parked sales auto-release inventory and log reason
1.1.2 Reports: Core Sales
| Report | Purpose | Key Data Fields |
|---|---|---|
| Daily Sales Summary | End-of-day overview of all transactions | Date, total sales, transaction count, avg transaction value, payment method breakdown |
| Item Entry Method Report | Track how items are entered into the system | Scanner vs manual entry counts, scanner success rate, failed scan count |
| Parked Sales Report | Monitor parked sales activity and expirations | Parked count, retrieved count, expired count, avg park duration |
| Hourly Sales Heatmap | Identify peak sales hours for staffing | Hour, transaction count, total revenue, avg items per transaction |
1.2 Discounts & Pricing Logic
Scope: Line-item overrides, Global discounts, Promotion application, Price Tiers, and Loyalty Redemptions.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Phase 2: Modification & Pricing
loop Pricing Actions
alt Line Item Modification
U->>UI: Select Item -> Override Price
opt Manager Auth
UI-->>U: Prompt PIN
U->>UI: Enter PIN
end
UI->>UI: Apply New Price & Reason Code (e.g., "Damaged")
end
alt Discounting
U->>UI: Apply Global Discount / Promo Code / Coupon
UI->>API: Validate Code (Expiry/Stacking/Single-Use)
API-->>UI: Valid
Note right of UI: Critical Calculation Order:
UI->>UI: 1. Apply Price Tier (Wholesale/VIP)
UI->>UI: 2. Apply Line Discounts
UI->>UI: 3. Apply Global % (Excluding Non-Discountable)
UI->>UI: 4. Apply Coupons
UI->>UI: 5. Calculate Tax (on discounted subtotal)
UI->>UI: 6. Apply Loyalty Redemptions (after tax)
end
end
1.2.1 Discount Calculation Order (Definitive)
The system applies discounts in this strict order:
| Order | Type | Example | Applies To |
|---|---|---|---|
| 1 | Price Tier | Wholesale pricing | Base price replacement |
| 2 | Line Discounts | “Damaged - 20% off” | Individual items |
| 3 | Automatic Promos | “Buy 2 Get 1 Free” | Qualifying items |
| 4 | Global Discount | “10% off entire order” | Subtotal (excl. non-discountable) |
| 5 | Coupons | “SAVE10” code | After global discount |
| 6 | Tax Calculation | State + Local tax | On final discounted amount |
| 7 | Loyalty Redemption | “500 points = $5 off” | Final subtotal after tax |
Non-Discountable Items: Gift cards, deposits, and items flagged is_discountable = false are excluded from global discounts.
1.2.2 Reports: Discounts & Pricing
| Report | Purpose | Key Data Fields |
|---|---|---|
| Discount Usage Report | Track all discounts applied | Discount type, frequency, total value, avg discount %, top discounted items |
| Promotion Effectiveness | Measure promo campaign success | Promo code, redemption count, revenue impact, items sold via promo |
| Coupon Performance | Track coupon redemption rates | Coupon code, issued count, redeemed count, expired count, revenue impact |
| Price Override Audit | Monitor manual price changes | Override count, avg override %, reason codes, authorizing manager |
1.3 Financial Settlement (Payments, Layaway, Third-Party Financing)
Scope: Split tenders, Credit Limits, Gift Cards, Third-Party Financing (Affirm), and Finalization.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
participant PAY as Payment Gateway
Note over U, DB: Phase 3: Settlement
U->>UI: Click "Pay"
loop Tender Processing
U->>UI: Select Method
Note right of UI: Multiple tenders allowed: cash + card(s), multiple cards, etc.
alt Cash
UI-->>U: Show "Collect: $20.03"
U->>UI: Enter Amount Received
UI->>UI: Calculate Change Due
else Gift Card
U->>UI: Scan/Enter Gift Card Number
UI->>API: GET /giftcards/{number}/balance
API-->>UI: Return Balance & Expiry
alt Sufficient Balance
UI->>UI: Deduct from Gift Card
UI->>UI: Add to Tender List
else Insufficient Balance
UI-->>U: "Card Balance: $X. Apply partial?"
U->>UI: Apply Partial Amount
end
else On-Account (Store Credit)
UI->>API: Check Credit Limit (Balance + Pending Layaways + Cart)
alt Exceeds Limit
API-->>UI: Block Transaction
else Approved
UI->>UI: Add to Tender List
end
else Layaway Deposit
U->>UI: Select Layaway Mode
UI->>UI: Validate Min Deposit %
Note right of UI: Sale Status -> LAYAWAY
else Credit Card (SAQ-A Flow)
UI->>API: POST /payments/initiate {amount, order_id}
API->>PAY: Send amount to terminal
Note over PAY: Customer taps/inserts card
PAY-->>API: Token + Approval Code
API->>DB: Store token (never card data)
API-->>UI: Payment Approved
else Third-Party Financing (Affirm)
U->>UI: Select "Pay with Affirm"
UI->>API: POST /payments/affirm/initiate {amount, order_id, customer}
API->>PAY: Create Affirm Checkout Session
PAY-->>API: Redirect URL / QR Code
API-->>UI: Display QR Code or Redirect
Note over PAY: Customer completes Affirm application on their device
PAY-->>API: Webhook: Loan Approved + Charge ID
API->>DB: Store Affirm charge_id, loan_id, status
API-->>UI: Payment Approved (Affirm)
Note right of API: Full amount received from Affirm; customer pays Affirm directly
end
UI->>UI: Update Remaining Balance
end
UI->>API: POST /orders/finalize
par Backend Operations
API->>DB: Write Order Record
API->>DB: Update Inventory (Decrement Sale / Increment Return)
API->>DB: Update Customer (Loyalty Points / Increase Debt)
API->>DB: Update Gift Card Balance (if used)
API->>DB: Record Commission (Employee ID + Amount)
end
API-->>UI: Transaction Success
opt Receipt Printing
U->>UI: Select Template (Thermal / A4 Invoice / Gift Receipt)
UI->>U: Print / Email Receipt
end
1.3.1 Order State Machine
stateDiagram-v2
[*] --> DRAFT: Cart Created
DRAFT --> PENDING: Click Pay
PENDING --> PARTIAL_PAID: Partial Payment
PARTIAL_PAID --> PENDING: More Payment Needed
PENDING --> PAID: Full Payment
PARTIAL_PAID --> PAID: Full Payment
PAID --> COMPLETED: Finalized
PAID --> HOLD_FOR_PICKUP: Hold Requested
HOLD_FOR_PICKUP --> READY_FOR_PICKUP: Items Staged
READY_FOR_PICKUP --> COMPLETED: Customer Picked Up
READY_FOR_PICKUP --> HOLD_EXPIRED: Deadline Passed
HOLD_EXPIRED --> CONTACT_CUSTOMER: Staff Notified
CONTACT_CUSTOMER --> READY_FOR_PICKUP: Deadline Extended
CONTACT_CUSTOMER --> REFUNDED: Customer Wants Refund
COMPLETED --> VOIDED: Void Action (Same Day)
COMPLETED --> PARTIALLY_RETURNED: Partial Return
PARTIALLY_RETURNED --> FULLY_RETURNED: All Items Returned
VOIDED --> [*]
FULLY_RETURNED --> [*]
REFUNDED --> [*]
1.3.2 Layaway State Machine
stateDiagram-v2
[*] --> DEPOSIT_PAID: Min Deposit Received
DEPOSIT_PAID --> RESERVED: Inventory Reserved
RESERVED --> RESERVED: Additional Payment
RESERVED --> PAID_IN_FULL: Final Payment
PAID_IN_FULL --> COMPLETED: Items Released
RESERVED --> CANCELLED: Customer Cancels
RESERVED --> FORFEITED: Payment Deadline Missed
CANCELLED --> [*]
FORFEITED --> [*]
COMPLETED --> [*]
1.3.3 Credit Limit Calculation
When checking if a customer can use On-Account payment:
Available Credit = Credit Limit - (Current Debt + Pending Layaway Balances + Current Cart Total)
Example:
- Credit Limit: $500
- Current Debt: $150
- Pending Layaway (remaining balance): $100
- Current Cart: $200
- Available Credit: $500 - ($150 + $100 + $200) = $50
- Result: Blocked (cart exceeds available credit by $150)
1.3.4 Reports: Financial Settlement
| Report | Purpose | Key Data Fields |
|---|---|---|
| Payment Method Breakdown | Analyze tender mix | Cash total, card total, gift card total, on-account total, Affirm total, split tender count |
| Affirm Financing Summary | Track third-party financing usage | Affirm transaction count, total financed, avg loan amount, approval rate |
| Layaway Status Report | Monitor active layaways | Active count, total deposits, total remaining balances, overdue count |
| On-Account Aging Report | Track customer credit usage | Customer, balance, credit limit, days outstanding, aging buckets (30/60/90) |
1.4 Post-Sale Management (History, Void, Returns, Exchanges)
Scope: Corrections, History, Returns, Dedicated Exchanges, Receipt Reprinting, and Data Export.
sequenceDiagram
autonumber
participant U as Manager
participant UI as POS UI
participant API as Backend
participant DB as DB
U->>UI: Open Sales History
U->>UI: Apply Filters (Date, User, Status)
UI->>API: GET /sales/history
alt Action: Void (Same-Day Correction Only)
Note right of U: "Oops, wrong item - same day"
UI->>API: GET /orders/{id}/void-eligibility
API-->>UI: Check if same business day & drawer still open
alt Eligible for Void
U->>UI: Click Void -> Confirm
UI->>API: POST /sales/{id}/void
par Reversal
API->>DB: Reverse Inventory & Loyalty
API->>DB: Reverse Commission (Full)
API->>DB: Set Status "VOIDED"
end
UI-->>U: Alert: "Manually Refund Card Terminal"
else Not Eligible (Different Day)
UI-->>U: "Cannot void - use Return instead"
end
else Action: Return (with Policy Check)
Note right of U: "Customer brought it back"
U->>UI: Scan Receipt Barcode
UI->>API: POST /receipts/validate {barcode_data}
API->>DB: Verify receipt authenticity & match to order
alt Receipt Valid
API-->>UI: Receipt Verified - Load Original Sale
else Receipt Invalid
API-->>UI: "Invalid Receipt - Cannot Process Return"
UI-->>U: "Receipt validation failed. Verify receipt."
end
UI->>API: GET /sales/{id}/return-eligibility
API-->>UI: Return Policy Result
alt Within Policy (Receipt + Time)
UI->>UI: Select Items to Return
UI->>UI: Process Refund to Original Payment
API->>DB: Reverse Commission (Proportional)
else Outside Policy (No Receipt / Expired)
UI-->>U: "Policy Exception - Store Credit Only"
opt Manager Override
UI-->>U: Prompt Manager PIN
U->>UI: Enter PIN + Reason Code
end
UI->>UI: Issue Store Credit
end
else Action: Exchange (Dedicated Flow)
Note right of U: "Customer wants different size"
U->>UI: Load Original Sale -> Click "Exchange"
UI->>UI: Select Item(s) to Exchange OUT
UI->>UI: Scan/Add New Item(s) IN
UI->>UI: Calculate Price Difference
alt Customer Owes Money
UI-->>U: "Collect: $15.00 difference"
U->>UI: Process Payment
else Store Owes Refund
UI-->>U: "Refund: $10.00 to customer"
U->>UI: Process Refund
else Even Exchange
UI->>UI: No Payment Required
end
UI->>API: POST /sales/exchange
API->>DB: Create Exchange Record (Links Old & New)
API->>DB: Update Inventory (Both Directions)
API->>DB: Adjust Commission (if price difference)
else Action: Reprint Receipt
U->>UI: Click "Reprint Receipt"
UI->>API: GET /orders/{id}/receipt
API-->>UI: Return Receipt Data
U->>UI: Select Format (Thermal / A4 / Email)
opt Email to Different Address
U->>UI: Enter Email Address
UI->>API: POST /orders/{id}/email-receipt
end
else Action: Pay Off Layaway
U->>UI: Retrieve Layaway
UI->>UI: Pay Remaining Balance
UI->>API: Finalize (Status: COMPLETED)
API->>DB: Release Reserve -> Sold
end
Cross-Reference: See Module 4, Section 4.13 for POS inventory integration details.
1.4.1 Void vs. Return Rules
| Aspect | Void | Return |
|---|---|---|
| When allowed | Same business day, drawer open | Any time within policy |
| Inventory | Reversed immediately | Reversed on completion |
| Commission | Full reversal | Proportional reversal |
| Card refund | Manual on terminal only | Staff chooses: Manual on terminal OR Automatic via token |
| Cash refund | Cash returned from drawer | Cash returned from drawer |
| Audit trail | “VOIDED” status | “RETURNED” line items |
| Use case | Cashier mistake | Customer return |
Card Refund Method Selection (Returns Only):
- Manual on terminal: Customer presents physical card. Staff processes refund on the payment terminal. Use when customer is present with their card.
- Automatic via token: System uses the stored payment token to process refund without card present. Use when customer does not have their card or for remote returns.
- Cash payments: Refund is always cash from drawer. No token option available for cash transactions.
1.4.2 Reports: Post-Sale
| Report | Purpose | Key Data Fields |
|---|---|---|
| Void Summary | Track voided transactions | Void count, total value, reason codes, voiding employee, authorizing manager |
| Return Summary | Track returns by period | Return count, total refund value, refund method breakdown, top returned items |
| Exchange Summary | Track exchange activity | Exchange count, price difference (net), upgrade vs downgrade ratio |
| Refund Method Report | Analyze refund processing | Manual on terminal count, automatic via token count, cash refund count |
Email Template: TMPL-REFUND-CONFIRMATION
| Field | Value |
|---|---|
| Trigger | Refund processed (card or store credit) |
| Recipient | Customer (if email on file) |
| Content | Refund amount, method, expected processing time, original order reference |
1.5 Gift Card Management
Scope: Selling, Activating, Redeeming, Balance Management, and Jurisdiction Compliance.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Gift Card Operations
alt Sell New Gift Card
U->>UI: Scan Gift Card (or Enter Number)
U->>UI: Enter Load Amount
UI->>API: POST /giftcards/activate
API->>DB: Create Gift Card Record (Number, Balance, Expiry)
API->>DB: Set Jurisdiction Rules (based on store location)
API-->>UI: Activation Success
UI->>UI: Add Gift Card to Cart as Product
Note right of UI: Proceeds through normal checkout
end
alt Check Balance
U->>UI: Click "Gift Card Balance"
U->>UI: Scan/Enter Card Number
UI->>API: GET /giftcards/{number}/balance
API-->>UI: Display Balance & Expiry Date (if applicable)
end
alt Reload Existing Card
U->>UI: Scan Gift Card
UI->>API: GET /giftcards/{number}
API-->>UI: Card Found - Current Balance
U->>UI: Enter Reload Amount
UI->>UI: Add Reload to Cart
Note right of UI: Balance updated after payment
end
alt Cash Out (California Compliance)
U->>UI: Scan Gift Card
UI->>API: GET /giftcards/{number}
API-->>UI: Return Balance & Jurisdiction
alt Balance <= Cash Out Threshold
UI-->>U: "Eligible for Cash Out: $8.50"
U->>UI: Process Cash Out
API->>DB: Zero Balance, Record Cash Out
else Balance > Threshold
UI-->>U: "Not eligible for cash out"
end
end
1.5.1 Gift Card State Machine
stateDiagram-v2
[*] --> INACTIVE: Card Manufactured
INACTIVE --> ACTIVE: Activated at POS (Sold)
ACTIVE --> ACTIVE: Partial Redemption
ACTIVE --> ACTIVE: Reload
ACTIVE --> DEPLETED: Balance = $0.00
ACTIVE --> EXPIRED: Past Expiry Date (where allowed)
DEPLETED --> ACTIVE: Reloaded
DEPLETED --> CASHED_OUT: Cash Out Processed
EXPIRED --> [*]
CASHED_OUT --> [*]
note right of ACTIVE
Balance > $0
Within expiry (if applicable)
end note
note right of DEPLETED
Balance = $0
Can be reloaded
end note
1.5.2 Gift Card Jurisdiction Compliance
| Rule | Virginia | California | New York | Default |
|---|---|---|---|---|
| Expiry Allowed | Yes (5yr min) | No | No | No |
| Inactivity Fees | Yes (after 12mo) | No | No | No |
| Cash Out Threshold | None | $10.00 | None | None |
| Cash Out Required | No | Yes | No | No |
Implementation: Store location determines which jurisdiction rules apply. System defaults to most restrictive (California-style) and enables features where permitted.
1.5.3 Reports: Gift Cards
| Report | Purpose | Key Data Fields |
|---|---|---|
| Gift Card Liability Report | Track outstanding gift card balances | Total active cards, total outstanding balance, avg balance per card |
| Gift Card Activity Report | Monitor gift card transactions | Activations, reloads, redemptions, cash-outs, expired cards by period |
| Gift Card Aging | Identify dormant cards | Card number, last activity date, balance, days inactive |
1.6 Special Orders & Back Orders
Scope: Ordering out-of-stock items with customer deposits.
sequenceDiagram
autonumber
participant U as Staff
participant C as Customer
participant UI as POS UI
participant API as Backend
participant DB as DB
participant INV as Inventory/Purchasing
Note over U, INV: Special Order Flow
U->>UI: Search Product -> Out of Stock
UI-->>U: "Available for Special Order"
U->>UI: Click "Create Special Order"
UI->>UI: Attach Customer (Required)
UI->>UI: Enter Quantity Needed
UI->>UI: Calculate Deposit (Min 50% or Full)
C->>U: Agrees to Deposit
U->>UI: Process Deposit Payment
UI->>API: POST /special-orders/create
par Order Creation
API->>DB: Create Special Order Record
API->>DB: Link Customer & Deposit Payment
API->>DB: Set Status: DEPOSIT_PAID
API->>INV: Notify Purchasing Team
end
API-->>UI: Order #SO-12345 Created
UI->>U: Print Special Order Receipt
Note over INV, DB: Later - Item Arrives
INV->>API: POST /special-orders/{id}/received
API->>DB: Update Status: READY_FOR_PICKUP
API-->>C: Send Notification (SMS/Email)
Note over U, C: Customer Pickup
U->>UI: Retrieve Special Order
UI->>UI: Show Deposit Already Paid
UI->>UI: Calculate Remaining Balance
U->>UI: Collect Remaining Payment
UI->>API: POST /special-orders/{id}/complete
API->>DB: Status: COMPLETED
1.6.1 Special Order State Machine
stateDiagram-v2
[*] --> CREATED: Order Initiated
CREATED --> DEPOSIT_PAID: Deposit Received
DEPOSIT_PAID --> ORDERED: Sent to Vendor
ORDERED --> RECEIVED: Item Arrived at Store
RECEIVED --> READY_FOR_PICKUP: Inspected & Staged
READY_FOR_PICKUP --> COMPLETED: Customer Picked Up
READY_FOR_PICKUP --> ABANDONED: No Pickup (30+ days)
CREATED --> CANCELLED: Customer Cancels (No Deposit)
DEPOSIT_PAID --> CANCELLED_REFUND: Customer Cancels (Refund Deposit)
CANCELLED --> [*]
CANCELLED_REFUND --> [*]
COMPLETED --> [*]
ABANDONED --> [*]
1.6.2 Reports: Special Orders
| Report | Purpose | Key Data Fields |
|---|---|---|
| Special Order Status Report | Track all special orders by status | Order ID, customer, item, status, deposit amount, days in current status |
| Special Order Aging | Identify overdue or stalled orders | Orders past expected date, orders approaching abandonment threshold |
Email Template: TMPL-SPECIAL-ORDER-READY
| Field | Value |
|---|---|
| Trigger | Special order status changes to READY_FOR_PICKUP |
| Recipient | Customer |
| Content | Order details, remaining balance, store address, pickup deadline |
1.7 Multi-Store Inventory & Transfers
Scope: Cross-store inventory lookup, transfers, and reservations (requires full payment).
Cross-Reference: See Module 4, Section 4.8 for inter-store transfer details.
sequenceDiagram
autonumber
participant U as Staff
participant C as Customer
participant UI as POS UI
participant API as Backend
participant DB as DB
participant S2 as Store B
Note over U, S2: Multi-Store Inventory Check
U->>UI: Search Product -> Low/No Stock
U->>UI: Click "Check Other Stores"
UI->>API: GET /inventory/multi-store/{sku}
Note right of API: Eventually consistent (max 5 min stale)
API-->>UI: Return Stock Levels (All Locations)
UI-->>U: Display: "Store B: 5 units, Store C: 2 units"
alt Request Transfer to This Store
U->>UI: Select Store B -> "Request Transfer"
UI->>UI: Enter Quantity
UI-->>U: "Customer must pay in full to process"
C->>U: Agrees to Pay
U->>UI: Add Item to Cart (Status: TRANSFER_PENDING)
U->>UI: Complete Full Payment
UI->>API: POST /transfers/request
API->>DB: Create Transfer Record (PAID)
API->>S2: Notify Store B to Ship
API-->>UI: Transfer #TRF-789 Created
UI-->>U: "Item will arrive in 2-3 days"
UI->>U: Print Transfer Receipt for Customer
else Reserve at Other Store (Customer Pickup)
U->>UI: Select Store B -> "Reserve for Pickup"
UI-->>U: "Customer must pay in full to reserve"
C->>U: Agrees to Pay
U->>UI: Process Full Payment
UI->>API: POST /reservations/create
API->>DB: Create Reservation (PAID)
API->>S2: Reserve Item at Store B
API-->>UI: Reservation #RES-456 Created
UI-->>U: "Reserved at Store B until [date]"
UI->>U: Print Pickup Voucher for Customer
end
Note over S2, DB: Store B Fulfillment
S2->>API: POST /transfers/{id}/picked
API->>DB: Update Status: PICKING
S2->>API: POST /transfers/{id}/shipped
API->>DB: Update Status: SHIPPED
Note right of DB: Carrier scan triggers IN_TRANSIT
API-->>U: Notification: "Transfer Shipped"
1.7.1 Transfer State Machine
stateDiagram-v2
[*] --> REQUESTED: Transfer Initiated
REQUESTED --> PAID: Customer Paid in Full
PAID --> PICKING: Source Store Processing
PICKING --> SHIPPED: Handed to Carrier
SHIPPED --> IN_TRANSIT: Carrier Scan Confirmed
IN_TRANSIT --> RECEIVED: Arrived at Destination
RECEIVED --> COMPLETED: Customer Notified & Picked Up
REQUESTED --> CANCELLED: Cancelled Before Payment
PAID --> CANCELLED_REFUND: Cancelled After Payment
CANCELLED --> [*]
CANCELLED_REFUND --> [*]
COMPLETED --> [*]
1.7.2 Reservation State Machine
stateDiagram-v2
[*] --> REQUESTED: Reservation Initiated
REQUESTED --> PAID: Customer Paid in Full
PAID --> RESERVED: Item Held at Store
RESERVED --> PICKED_UP: Customer Collected
RESERVED --> EXPIRED: Reservation Deadline Passed
EXPIRED --> REFUND_PENDING: Auto-Refund Triggered
REFUND_PENDING --> REFUNDED: Refund Processed
REQUESTED --> CANCELLED: Cancelled Before Payment
CANCELLED --> [*]
PICKED_UP --> [*]
REFUNDED --> [*]
1.7.3 Ship to Customer from Other Location
Scope: Direct shipping from a source store to the customer’s address, with carrier integration for real-time shipping cost calculation.
sequenceDiagram
autonumber
participant U as Staff
participant C as Customer
participant UI as POS UI
participant API as Backend
participant DB as DB
participant S2 as Source Store
participant SHIP as Carrier API
Note over U, SHIP: Ship to Customer from Another Store
U->>UI: Search Product -> Low/No Stock
U->>UI: Click "Check Other Stores"
UI->>API: GET /inventory/multi-store/{sku}
API-->>UI: Return Stock Levels (All Locations)
U->>UI: Select Source Store -> "Ship to Customer"
UI-->>U: "Enter Customer Shipping Address"
C->>U: Provides Shipping Address
U->>UI: Enter Address Details
UI->>API: POST /shipping/calculate
Note right of API: {origin_store, destination_address, items, weight}
API->>SHIP: Request Shipping Rates
SHIP-->>API: Return Shipping Options
API-->>UI: Display Shipping Options
UI-->>U: "Standard (3-5 days): $8.99 | Express (1-2 days): $15.99"
C->>U: Selects Shipping Option
U->>UI: Add Item + Shipping to Cart
UI->>UI: Total = Item Price + Shipping Cost
UI-->>U: "Customer must pay in full"
C->>U: Pays Full Amount (Item + Shipping)
U->>UI: Process Payment
UI->>API: POST /shipments/create
API->>DB: Create Shipment Record (PAID)
API->>S2: Notify Source Store to Pack & Ship
API-->>UI: Shipment #SHP-101 Created
UI-->>U: "Item will be shipped to customer"
UI->>U: Print Shipment Receipt for Customer
Note over S2, SHIP: Source Store Fulfillment
S2->>API: POST /shipments/{id}/packed
API->>DB: Update Status: PACKED
S2->>SHIP: Request Shipping Label
SHIP-->>S2: Return Label + Tracking Number
S2->>API: POST /shipments/{id}/shipped {tracking_number}
API->>DB: Update Status: SHIPPED
API-->>C: Send Tracking Email to Customer
Note over SHIP, DB: Delivery
SHIP->>API: Webhook: Delivered
API->>DB: Update Status: DELIVERED
API-->>C: Send Delivery Confirmation Email
1.7.4 Ship-to-Customer State Machine
stateDiagram-v2
[*] --> REQUESTED: Shipment Initiated
REQUESTED --> PAID: Customer Paid (Item + Shipping)
PAID --> PICKING: Source Store Processing
PICKING --> PACKED: Items Packed
PACKED --> SHIPPED: Label Generated & Handed to Carrier
SHIPPED --> IN_TRANSIT: Carrier Pickup Confirmed
IN_TRANSIT --> DELIVERED: Delivery Confirmed
REQUESTED --> CANCELLED: Cancelled Before Payment
PAID --> CANCELLED_REFUND: Cancelled After Payment (Full Refund)
CANCELLED --> [*]
CANCELLED_REFUND --> [*]
DELIVERED --> [*]
1.7.5 Reports: Multi-Store & Shipping
| Report | Purpose | Key Data Fields |
|---|---|---|
| Transfer Status Report | Track inter-store transfers | Transfer ID, source/destination, status, days in transit, customer |
| Shipping Fulfillment Report | Monitor ship-to-customer orders | Shipment ID, carrier, tracking, status, delivery date, shipping cost |
| Reservation Report | Track cross-store reservations | Reservation ID, store, item, status, expiry date, customer |
| Multi-Store Inventory Discrepancy | Flag stock mismatches after sync | SKU, expected vs actual, location, last sync time |
Email Template: TMPL-SHIPMENT-TRACKING
| Field | Value |
|---|---|
| Trigger | Shipment status changes to SHIPPED |
| Recipient | Customer |
| Content | Tracking number, carrier, estimated delivery, order details |
Email Template: TMPL-DELIVERY-CONFIRMATION
| Field | Value |
|---|---|
| Trigger | Shipment status changes to DELIVERED |
| Recipient | Customer |
| Content | Delivery confirmation, order summary, return/exchange policy link |
1.8 Sales Commissions
Scope: Track employee sales for commission calculation with proportional reversal on returns.
Cross-Reference: See Module 5, Section 5.5 for user commission rate configuration.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Commission Tracking (Per Transaction)
U->>UI: Login to POS (Employee ID Captured)
Note right of UI: Throughout Sale...
UI->>UI: Employee ID attached to session
U->>UI: Complete Sale
UI->>API: POST /orders/finalize
par Commission Recording
API->>DB: Calculate Commission Amount
Note right of API: Based on: Sale Total, Product Categories, Employee Tier
API->>DB: Insert Commission Record
Note right of DB: {employee_id, order_id, amount, date, line_items[]}
end
Note over U, DB: Commission Reversal on Return
U->>UI: Process Return (2 of 3 items)
UI->>API: POST /returns/create
par Proportional Reversal
API->>DB: Calculate Return Value / Original Sale Value
Note right of API: $80 returned / $120 sale = 66.7%
API->>DB: Reverse 66.7% of Commission
API->>DB: Update Commission Record
end
Note over U, DB: Commission Reporting
U->>UI: Manager -> Reports -> Commissions
UI->>API: GET /reports/commissions?date_range&employee
API->>DB: Aggregate Commission Data
API-->>UI: Return Commission Summary
UI-->>U: Display: Employee | Sales | Returns | Net Commission
1.8.1 Commission Calculation Rules
commission_calculation:
# Base calculation
base_method: "percentage_of_sale"
# Reversal rules
void_reversal: "full" # 100% reversal on void
return_reversal: "proportional" # Based on returned value
# Proportional calculation
# Commission Adjustment = Original Commission × (Returned Value / Original Sale Value)
# Example:
# Original Sale: $120, Commission: $6.00 (5%)
# Return: $80 worth of items
# Reversal: $6.00 × ($80/$120) = $4.00
# Net Commission: $6.00 - $4.00 = $2.00
1.8.2 Reports: Commissions
| Report | Purpose | Key Data Fields |
|---|---|---|
| Commission Summary | Period overview of all commissions | Total sales, total commissions, avg commission rate, top earners |
| Commission by Employee | Individual employee performance | Employee, sales count, sales total, commission earned, returns impact |
| Commission Reversal Log | Track commission adjustments | Order ID, original commission, reversal amount, reversal type (void/return), date |
1.9 Return Policy Engine
Configuration Note: Store return and exchange policies are manually configured in the application’s Settings/Setup module. Policies are NOT hardcoded in the application. Each tenant can configure different policies per store location and per sales channel (online vs in-store).
Scope: Configurable return rules based on receipt, time, and item type.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Return Policy Evaluation
U->>UI: Start Return
alt Has Receipt
U->>UI: Scan Receipt Barcode
UI->>API: POST /receipts/validate {barcode_data}
API->>DB: Verify receipt authenticity & match to order
alt Receipt Valid
API-->>UI: Receipt Verified
else Receipt Invalid
API-->>UI: "Invalid Receipt"
UI-->>U: "Receipt validation failed"
end
UI->>API: GET /orders/{id}
API-->>UI: Return Original Sale Data
UI->>UI: Calculate Days Since Purchase
alt Within 30 Days
UI-->>U: "Full Refund Eligible"
UI->>UI: Refund to Original Payment Method
else 31-90 Days
UI-->>U: "Store Credit Only"
UI->>UI: Issue Store Credit
else Beyond 90 Days
UI-->>U: "Manager Approval Required"
U->>UI: Enter Manager PIN
U->>UI: Select Exception Reason
end
else No Receipt
U->>UI: Scan Item Barcode
UI->>API: GET /products/{sku}/current-price
API-->>UI: Return Current Selling Price
UI-->>U: "No Receipt - Store Credit at Current Price"
UI-->>U: "Manager Approval Required"
U->>UI: Enter Manager PIN
UI->>UI: Issue Store Credit (Current Price)
end
alt Item-Specific Rules
opt Final Sale Item
UI-->>U: "BLOCKED: Final Sale - No Returns"
end
opt Opened Electronics
UI-->>U: "Restocking Fee: 15%"
UI->>UI: Deduct Restocking Fee
end
end
1.9.1 Default Policy Configuration Examples
Note: These are example default values only. Actual policies are configured per tenant in the application’s Settings/Setup module and are NOT hardcoded.
Online Sales Policy
| Policy | Timeframe | Refund Method | Conditions |
|---|---|---|---|
| Return | 30 days from delivery | Original payment method minus shipping & processing fees | Item in original condition, receipt required |
| Exchange | 30 days from delivery | Price difference applies; shipping & processing fees excluded from refund | Same category items preferred |
In-Store Sales Policy
| Policy | Timeframe | Refund Method | Conditions |
|---|---|---|---|
| Return | 24 hours from purchase | Original payment method | Receipt required (scanned for validation) |
| Exchange | 24 hours from purchase | Price difference applies | Receipt required (scanned for validation) |
Policy Configuration Fields (Settings/Setup):
- Return window (days/hours per channel)
- Exchange window (days/hours per channel)
- Refund method options (original method, store credit, etc.)
- Restocking fee percentage and applicable categories
- Final sale categories
- Manager override permissions
- Online shipping/processing fee exclusion rules
1.9.2 Reports: Return Policy
| Report | Purpose | Key Data Fields |
|---|---|---|
| Return Policy Exception Report | Track manager overrides on return policy | Order ID, exception type, reason code, authorizing manager, refund amount |
| Return Reason Analysis | Understand why customers return items | Reason code, frequency, product categories, avg refund value |
| Channel Return Comparison | Compare online vs in-store returns | Channel, return count, return rate, avg processing time |
1.10 Serial Number Tracking
Scope: Capture serial numbers for designated high-value items.
Cross-Reference: See Module 4, Section 4.10 for serial number tracking lifecycle.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Serial Number Capture at Sale
U->>UI: Scan Product (Serial Required Flag)
UI-->>U: "Enter Serial Number"
U->>UI: Scan/Enter Serial Number
UI->>API: POST /serials/validate
alt Serial Already Sold
API-->>UI: "ERROR: Serial already in system"
UI-->>U: Block Item - Investigate
else Serial Valid/New
API-->>UI: Valid
UI->>UI: Attach Serial to Line Item
end
U->>UI: Complete Sale
UI->>API: POST /orders/finalize
API->>DB: Store Serial with Order Line
Note right of DB: {order_id, line_id, serial_number, product_sku}
Note over U, DB: Serial Lookup (Returns/Warranty)
U->>UI: Search by Serial Number
UI->>API: GET /serials/{number}
API-->>UI: Return Purchase History
UI-->>U: "Sold on [date] to [customer] - Order #123"
1.10.1 Reports: Serial Number Tracking
| Report | Purpose | Key Data Fields |
|---|---|---|
| Serial Number Audit Trail | Full history of serial-tracked items | Serial number, product SKU, sale date, customer, return status |
| Missing Serial Report | Flag transactions missing required serials | Order ID, product, serial required flag, serial captured (Y/N) |
1.11 Hold for Pickup (Including BOPIS)
Scope: Fully paid items held for customer pickup at the CURRENT store. This includes both in-store holds and BOPIS (Buy Online, Pick Up In Store) orders. Different from Layaway (partial payment) and Reservation (item at another store).
Cross-Reference: See Module 4, Section 4.13 for inventory reservation model.
Reservation vs Hold for Pickup - Key Distinction:
Aspect Reservation (Section 1.7.2) Hold for Pickup (Section 1.11) What it is Reserve item at a DIFFERENT store for customer pickup Pay for items at THIS store, pick up later Origin In-store POS (staff-initiated) In-store POS or Online order (BOPIS) Payment Full payment at originating store Full payment required upfront Inventory location At the remote store At the current store Customer picks up at The remote store The same store (or originating store for BOPIS) BOPIS No Yes - this is the BOPIS flow Use case “Store B has it, I’ll drive there to get it” “I’ll pay now, come back Saturday” or “Order online, pick up in store” Examples:
Reservation Example: Customer is at Store A. Item is out of stock. Store B has 3 units. Customer pays at Store A. Item is reserved at Store B. Customer drives to Store B to pick it up with their pickup voucher.
Hold for Pickup Example 1 (In-Store): Customer at Store A buys a large piece of furniture. Pays in full. Asks store to hold it until Saturday when they can bring a truck. Store stages the item. Customer returns Saturday to pick up.
Hold for Pickup Example 2 (Online/BOPIS): Customer browses the online store. Selects “Pick Up at Store A.” Pays online. Store A receives the order, stages the items. Customer receives “Ready for Pickup” notification. Customer walks into Store A and picks up.
sequenceDiagram
autonumber
participant U as Staff
participant C as Customer
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Hold for Pickup Flow
U->>UI: Add Items to Cart
U->>UI: Attach Customer
U->>UI: Click "Hold for Pickup"
UI->>UI: Set Pickup Deadline (Default: 7 days)
UI-->>U: "Customer must pay in full"
C->>U: Pays Full Amount
U->>UI: Process Payment
UI->>API: POST /orders/finalize
API->>DB: Create Order (Status: HOLD_FOR_PICKUP)
API->>DB: Set Pickup Deadline
API->>DB: Reserve Inventory
API-->>UI: Order Complete
UI->>U: Print Pickup Slip
Note over U, DB: Customer Returns to Pickup
U->>UI: Retrieve Held Order
UI->>API: GET /orders/held/{id}
API-->>UI: Display Order Details
U->>UI: Verify Customer ID
U->>UI: Click "Release to Customer"
UI->>API: PATCH /orders/{id}/pickup-complete
API->>DB: Status: COMPLETED
API->>DB: Release Inventory Hold
Note over API, DB: Expiry Handling (Background Job)
API->>DB: Check Overdue Holds Daily
alt Hold Expired
API->>DB: Status: HOLD_EXPIRED
API-->>U: Alert: "Hold #123 expired - contact customer"
end
1.11.1 Hold for Pickup State Machine
stateDiagram-v2
[*] --> HOLD_FOR_PICKUP: Full Payment + Hold Request
HOLD_FOR_PICKUP --> READY_FOR_PICKUP: Items Staged
READY_FOR_PICKUP --> COMPLETED: Customer Picked Up
READY_FOR_PICKUP --> HOLD_EXPIRED: Deadline Passed
HOLD_EXPIRED --> CONTACT_CUSTOMER: Staff Notified
CONTACT_CUSTOMER --> READY_FOR_PICKUP: Deadline Extended
CONTACT_CUSTOMER --> REFUNDED: Customer Wants Refund
REFUNDED --> [*]
COMPLETED --> [*]
1.12 Cash Drawer Management
Scope: Opening float, blind counts, variance tracking, X-reports (mid-shift), and Z-reports (end-of-shift).
Cross-Reference: See Module 5, Section 5.7 for register configuration and Section 5.6 for clock-in/clock-out time tracking.
sequenceDiagram
autonumber
participant U as Staff
participant M as Manager
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Shift Start - Open Drawer
M->>UI: Open Cash Drawer Session
M->>UI: Enter Opening Float Amount
UI->>API: POST /cash-drawer/open
API->>DB: Create Drawer Session (opening_float, start_time)
API-->>UI: Drawer Session Started
Note over U, DB: During Shift - Transactions
loop Cash Transactions
U->>UI: Process Cash Sale/Refund
UI->>DB: Record Cash In/Out
end
opt Mid-Shift Check (X-Report)
U->>UI: Click "X-Report"
UI->>API: GET /cash-drawer/x-report
API->>DB: Calculate Current Expected Amount
Note right of API: Expected = Float + Cash Sales So Far - Cash Refunds - Payouts
API-->>UI: Return X-Report Data
UI->>U: Print/Display X-Report
Note right of UI: X-Report does NOT close the drawer
Note right of UI: Can be run multiple times per shift
end
Note over U, DB: Shift End - Close Drawer
U->>UI: Click "Close Drawer"
UI-->>U: "Perform Blind Count"
U->>UI: Enter Counted Cash (Blind - no expected shown)
UI->>API: POST /cash-drawer/count
API->>DB: Calculate Expected Amount
Note right of API: Expected = Float + Cash Sales - Cash Refunds - Payouts
API->>DB: Calculate Variance (Counted - Expected)
alt Variance Within Tolerance
API-->>UI: "Drawer Balanced"
else Variance Outside Tolerance
API-->>UI: "Variance: -$5.00 - Manager Approval Required"
M->>UI: Enter PIN + Variance Reason
end
API->>DB: Close Drawer Session
API->>DB: Record Final Counts & Variance
UI->>U: Print Z-Report
Note over M, DB: Z-Report Contents
Note right of UI: - Opening Float
Note right of UI: - Cash Sales Total
Note right of UI: - Cash Refunds Total
Note right of UI: - Expected Cash
Note right of UI: - Counted Cash
Note right of UI: - Variance (+/-)
Note right of UI: - Transaction Count
1.12.1 Cash Drawer State Machine
stateDiagram-v2
[*] --> CLOSED: Drawer Secured
CLOSED --> OPENING: Manager Opens
OPENING --> OPEN: Float Entered
OPEN --> OPEN: Transactions Processed
OPEN --> COUNTING: Close Initiated
COUNTING --> BALANCED: Variance Within Tolerance
COUNTING --> VARIANCE_DETECTED: Variance Outside Tolerance
VARIANCE_DETECTED --> MANAGER_REVIEW: Awaiting Approval
MANAGER_REVIEW --> BALANCED: Manager Approved
BALANCED --> CLOSED: Z-Report Printed
CLOSED --> [*]
1.12.2 X-Report vs Z-Report
| Aspect | X-Report | Z-Report |
|---|---|---|
| When | Any time during shift | End of shift only |
| Closes Drawer | No | Yes |
| Resets Counters | No | Yes |
| Frequency | Unlimited per shift | Once per shift |
| Use Cases | Mid-shift audit, shift handoff check, manager spot-check | End-of-day close, final reconciliation |
| Content | Same as Z-Report (opening float, sales, refunds, expected cash) | Same content + final blind count + variance |
X-Report Use Cases:
- Manager wants to verify cash during a busy period
- Shift handoff between employees (outgoing staff checks before handing off)
- Routine mid-day audit required by store policy
- Investigating a suspected cash handling issue
1.12.3 Reports: Cash Drawer
| Report | Purpose | Key Data Fields |
|---|---|---|
| X-Report | Mid-shift cash snapshot (does not close drawer) | Opening float, cash sales, cash refunds, payouts, expected cash |
| Z-Report | End-of-shift final reconciliation | Same as X-Report + blind count, variance, manager approval |
| Variance History Report | Track cash variances over time | Date, shift, employee, expected, counted, variance, reason code |
| Cash Movement Log | Detailed cash in/out record | Timestamp, type (sale/refund/payout/float), amount, employee |
1.13 Price Check Mode
Scope: Quick price lookup without adding to cart.
sequenceDiagram
autonumber
participant U as Staff
participant C as Customer
participant UI as POS UI
participant API as Backend
C->>U: "How much is this?"
U->>UI: Click "Price Check" Mode
UI->>UI: Switch to Price Check Display
U->>UI: Scan Item Barcode
UI->>API: GET /products/{sku}
API-->>UI: Return Product Info
UI-->>U: Display Large Price
UI-->>U: Show: Name, SKU, Price, Stock Level
opt Promotion Active
UI-->>U: "ON SALE: Was $50, Now $39.99"
end
U->>UI: Press Any Key / Timeout
UI->>UI: Return to Normal Sale Mode
1.14 Coupon System
Scope: Single-use and multi-use coupons (separate from promo codes).
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Coupon Application
U->>UI: Click "Apply Coupon"
U->>UI: Scan/Enter Coupon Code
UI->>API: POST /coupons/validate
API->>DB: Lookup Coupon
alt Single-Use Coupon (e.g., Birthday)
API->>DB: Check if Already Redeemed
alt Already Used
API-->>UI: "Coupon Already Redeemed"
else Valid
API-->>UI: Return Discount Details
end
else Multi-Use Coupon (e.g., SAVE10)
API->>DB: Check Usage Limit & Expiry
API-->>UI: Return Discount Details
end
alt Valid Coupon
UI->>UI: Apply Discount
UI->>UI: Display Savings
Note right of UI: Coupon marked for redemption at finalize
end
U->>UI: Complete Sale
UI->>API: POST /orders/finalize
par Coupon Processing
API->>DB: Mark Single-Use Coupon as REDEEMED
API->>DB: Increment Multi-Use Coupon Counter
API->>DB: Link Coupon to Order
end
1.14.1 Coupon State Machine
stateDiagram-v2
[*] --> CREATED: Coupon Generated
CREATED --> ACTIVE: Published/Distributed
state ACTIVE {
[*] --> AVAILABLE
AVAILABLE --> APPLIED: Added to Cart
APPLIED --> AVAILABLE: Removed from Cart
APPLIED --> REDEEMED: Order Finalized
}
ACTIVE --> EXPIRED: Past Expiry Date
ACTIVE --> DEPLETED: Usage Limit Reached (Multi-Use)
REDEEMED --> [*]: Single-Use Complete
EXPIRED --> [*]
DEPLETED --> [*]
1.15 Flexible Loyalty Programs
Scope: Configurable loyalty: points-per-dollar, punch cards, spend thresholds.
Cross-Reference: See Module 5, Section 5.17 for loyalty program settings configuration (point rates, tier thresholds, gift card settings).
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Loyalty Program Types
U->>UI: Attach Customer to Sale
UI->>API: GET /customers/{id}/loyalty
API-->>UI: Return Loyalty Profile & Active Programs
alt Points Per Dollar Program
Note right of UI: Earn 1 point per $1 spent
UI->>UI: Calculate Points to Earn
UI-->>U: "Customer earns 45 points"
opt Redeem Points
U->>UI: Click "Redeem Points"
UI-->>U: "500 points = $5 off"
U->>UI: Apply Redemption
end
else Punch Card Program
Note right of UI: Buy 10, Get 1 Free
UI->>UI: Check Qualifying Items in Cart
UI-->>U: "Coffee Purchase: Punch 3 of 10"
opt Card Complete
UI-->>U: "FREE ITEM EARNED!"
UI->>UI: Auto-Apply Free Item Discount
end
else Spend Threshold Program
Note right of UI: Spend $100, Get $10 Off
UI->>UI: Check Customer's Period Spend
UI-->>U: "Customer has spent $85 this month"
opt Threshold Reached This Sale
UI-->>U: "$10 Reward Unlocked!"
UI->>UI: Apply or Save for Next Visit
end
end
U->>UI: Complete Sale
UI->>API: POST /orders/finalize
par Loyalty Updates
API->>DB: Award Points Earned
API->>DB: Update Punch Card Count
API->>DB: Update Spend Totals
API->>DB: Check Tier Upgrades
end
1.15.1 Customer Tier State Machine
stateDiagram-v2
[*] --> BRONZE: New Customer
BRONZE --> SILVER: Spend >= $1,000/year
SILVER --> GOLD: Spend >= $5,000/year
GOLD --> GOLD: Maintains Spend
GOLD --> SILVER: Annual Spend < $5,000
SILVER --> BRONZE: Annual Spend < $1,000
note right of BRONZE
1x points
Standard pricing
end note
note right of SILVER
1.5x points
5% discount
end note
note right of GOLD
2x points
10% discount
Early access
end note
1.15.2 Reports: Loyalty Programs
| Report | Purpose | Key Data Fields |
|---|---|---|
| Loyalty Points Summary | Track points economy | Total points issued, redeemed, expired, outstanding balance |
| Tier Distribution Report | Customer breakdown by tier | Tier, customer count, avg spend, upgrade/downgrade count |
| Points Expiry Forecast | Predict upcoming point expirations | Customer, points expiring, expiry date, days remaining |
| Punch Card Activity | Track punch card completions | Program, cards started, cards completed, avg completion time |
| Loyalty ROI Analysis | Measure loyalty program value | Points cost, additional revenue from loyalty customers, retention rate |
1.16 Offline Operations
Scope: Queue-and-sync architecture for network resilience.
Cross-Reference: See Module 4, Section 4.15 for offline inventory operations.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant Q as Local Queue
participant API as Backend
participant DB as DB
Note over U, DB: Network Loss Detected
UI->>UI: Detect Network Unavailable
UI-->>U: Display "OFFLINE MODE" Indicator
Note over U, Q: Processing Sales Offline
U->>UI: Process Sale
UI->>Q: Queue Transaction Locally
Note right of Q: Store: items, customer_id, payments, timestamp
Q-->>UI: Transaction Queued (#1)
UI-->>U: "Sale Complete (Pending Sync)"
UI->>U: Print Receipt with "OFFLINE" watermark
loop Continue Offline Sales
U->>UI: Process More Sales
UI->>Q: Queue Each Transaction
UI-->>U: Show Queue Count (e.g., "3 pending")
end
Note over UI, DB: Network Restored
UI->>UI: Detect Network Available
UI-->>U: Display "SYNCING..." Indicator
loop Sync Queue
Q->>API: POST /orders/sync
API->>DB: Validate & Write
alt Conflict: Item Out of Stock
API-->>UI: Conflict Detected
UI-->>U: Alert: "Item X sold out - review order #1"
Note right of UI: Flag for manager review
else Success
API-->>UI: Synced
Q->>Q: Remove from Queue
end
end
UI-->>U: "All transactions synced"
1.16.1 Offline Mode State Machine
stateDiagram-v2
[*] --> ONLINE: Network Available
ONLINE --> OFFLINE: Network Lost
OFFLINE --> SYNCING: Network Restored
SYNCING --> ONLINE: Queue Empty
SYNCING --> CONFLICT_REVIEW: Conflicts Detected
CONFLICT_REVIEW --> SYNCING: Conflicts Resolved
CONFLICT_REVIEW --> ONLINE: Manager Override
note right of OFFLINE
Queue transactions locally
Limited operations only
end note
note right of SYNCING
Processing queue
Show progress
end note
1.16.2 Offline Operations Rules
offline_mode:
# Maximum transactions to queue locally
max_queue_size: 100
# Auto-sync interval when online (seconds)
sync_interval_seconds: 30
# Conflict resolution strategy
conflict_strategy: "server_wins_with_review"
# Operations ALLOWED offline
allowed_offline:
- sale_new
- return_with_receipt
- price_check
- gift_card_balance_check # Cached balance, show warning
- parked_sale_create
- parked_sale_retrieve
# Operations BLOCKED offline
blocked_offline:
- customer_create # Requires uniqueness check
- credit_limit_check # Requires real-time balance
- on_account_payment # Risk of exceeding limit
- multi_store_inventory # Requires network
- gift_card_activation # Must register immediately
- gift_card_reload # Risk of double-load
- transfer_request # Requires other store
- reservation_create # Requires other store
# Gift card handling offline
gift_card_offline:
balance_check: "use_cached" # Show cached with warning
redemption: "block" # Too risky - balance could be stale
# Inventory conflict handling
inventory_conflict:
# If offline sale sold item that's now out of stock
resolution: "manager_review"
auto_backorder: false
1.16.3 Offline Sync Conflict Resolution
| Conflict Type | Resolution | Action |
|---|---|---|
| Item out of stock | Manager review | Flag order, allow override or backorder |
| Price changed | Use offline price | Honor price at time of sale |
| Customer deleted | Use anonymous | Reassign to “Walk-in Customer” |
| Promotion expired | Manager review | Allow or remove discount |
| Gift card insufficient | Block sync | Require additional payment |
| Transfer item sold at source | Customer notification | Email customer (TMPL-OFFLINE-SOLD), ask to contact store to discuss options |
| Ship-to-customer item sold at source | Customer notification | Email customer (TMPL-OFFLINE-SOLD), ask to contact store to discuss options |
| Reservation item sold at source | Customer notification | Email customer (TMPL-OFFLINE-SOLD), ask to contact store to discuss options |
1.16.4 Offline Conflict Email Templates
When the system comes back online and discovers that a Transfer, Ship-to-Customer, or Reservation request cannot be fulfilled because the item was sold from the source location during the offline period, the system must automatically notify the customer.
Email Template: TMPL-OFFLINE-SOLD (Item Availability Change Notification)
| Field | Value |
|---|---|
| Template ID | TMPL-OFFLINE-SOLD |
| Trigger | Offline sync detects item sold from source location for Transfer, Ship-to-Customer, or Reserve requests |
| Recipient | Customer (email on file) |
| Subject | “Update Regarding Your Order - Action Required” |
| Content | Informs customer that the requested item is no longer available at the source location. Asks the customer to contact the store at their earliest convenience to discuss available options (alternative locations, backorder, refund, etc.) |
| Fallback | If no customer email on file, create staff alert for manual outreach |
Offline Conflict Notification Rules:
offline_conflict_notifications:
# Notify customer when item sold from source location
item_sold_at_source:
notify_customer: true
email_template: "TMPL-OFFLINE-SOLD"
fallback_if_no_email: "staff_alert"
# Staff alert appears in manager dashboard
staff_alert_priority: "high"
1.17 Tax Calculation Engine
Scope: Custom tax engine with jurisdiction support, hierarchy rules, and exemptions.
Cross-Reference: See Module 5, Section 5.9 for tax jurisdiction setup, compound rate configuration (State/County/City), and rate assignment per location.
sequenceDiagram
autonumber
participant UI as POS UI
participant API as Backend
participant TAX as Tax Engine
participant DB as DB
Note over UI, DB: Tax Calculation Flow
UI->>API: Calculate Tax for Cart
API->>TAX: Submit Cart + Store Location + Customer
TAX->>DB: Get Store Tax Jurisdiction
Note right of DB: Store in Virginia: State 4.3% + Local 1%
TAX->>DB: Get Customer Tax Status
Note right of DB: Customer: Regular (no exemption)
loop For Each Line Item
TAX->>DB: Get Product Tax Category
alt Product Override (e.g., Food)
TAX->>TAX: Apply Product Tax Rate (0% for groceries)
else Customer Exempt
TAX->>TAX: Apply 0% Tax
else Standard
TAX->>TAX: Apply Jurisdiction Rate (5.3%)
end
end
TAX-->>API: Return Tax Breakdown
API-->>UI: Display Tax Summary
Note right of UI: Tax Breakdown:
Note right of UI: State Tax (4.3%): $4.30
Note right of UI: Local Tax (1.0%): $1.00
Note right of UI: Total Tax: $5.30
1.17.1 Tax Hierarchy (Priority Order)
1. Product-Level Override (Highest Priority)
└── Example: "Grocery - Tax Exempt", "Prepared Food - 10%"
2. Customer-Level Exemption
└── Example: "Reseller Certificate", "Diplomatic Status", "Non-Profit"
3. Location-Based Tax (Default)
└── Based on store physical address
└── Includes: State + County + City + Special District
1.17.2 Virginia Tax Configuration
tax_jurisdictions:
virginia:
state_rate: 4.3
# Regional taxes (in addition to state)
regions:
hampton_roads:
counties: ["Norfolk", "Virginia Beach", "Newport News", "Hampton"]
additional_rate: 0.7
northern_virginia:
counties: ["Arlington", "Fairfax", "Loudoun", "Prince William"]
additional_rate: 0.7
central_virginia:
counties: ["Henrico", "Chesterfield", "Richmond City"]
additional_rate: 0.0
# Local tax (most localities)
default_local_rate: 1.0
# Product exemptions
exemptions:
- category: "grocery_food"
rate: 1.5 # Reduced rate for groceries
- category: "prescription_drugs"
rate: 0.0
- category: "medical_equipment"
rate: 0.0
tax_exemption_types:
- code: "RESALE"
description: "Reseller Certificate"
requires_certificate: true
certificate_expiry: true
- code: "NONPROFIT"
description: "501(c)(3) Non-Profit"
requires_certificate: true
certificate_expiry: true
- code: "DIPLOMAT"
description: "Diplomatic Exemption"
requires_certificate: true
certificate_expiry: false
- code: "NATIVE"
description: "Native American Tribal Member"
requires_certificate: true
certificate_expiry: false
1.17.3 Tax Engine Design for Expansion
Note: The Virginia compound tax jurisdiction model is the active reference implementation. See Section 5.9 for the
tax_jurisdictionsandtax_ratestables that implement the 3-level compound model (State + County + City).
jurisdiction_modules:
virginia:
status: "active" # Reference implementation
model: "compound_3_level" # State + County/Regional + City
sales_tax: true
use_tax: false
california:
status: "planned"
sales_tax: true
district_taxes: true # Complex district overlay
oregon:
status: "planned"
sales_tax: false # No sales tax state
canada:
status: "planned"
gst: true
pst: true # Varies by province
hst: true # Harmonized in some provinces
european_union:
status: "planned"
vat: true
reverse_charge: true # B2B cross-border
1.18 Payment Integration (SAQ-A)
Scope: Semi-integrated payment terminal architecture with PCI SAQ-A compliance.
Cross-Reference: Payment data storage rules, terminal communication protocol, processor configuration, and failure handling details have been consolidated into Module 6, Section 6.8. The payment flow sequence diagram below remains in this section as it is part of the core sales workflow.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Your Backend
participant TERM as Payment Terminal
participant PROC as Payment Processor
Note over U, PROC: Card Payment Flow (SAQ-A)
U->>UI: Click "Pay by Card"
UI->>API: POST /payments/initiate
Note right of API: {order_id, amount, terminal_id}
API->>TERM: Send Payment Request
Note right of TERM: Amount: $45.00
TERM-->>U: "Insert/Tap Card"
Note over TERM: Customer interacts with terminal only
U->>TERM: Customer taps card
TERM->>PROC: Encrypted Card Data
Note right of PROC: Card data NEVER touches your system
PROC-->>TERM: Authorization Response
TERM-->>API: Token + Approval Code + Masked Card
API->>API: Store Token (NOT card data)
Note right of API: Stored: token, approval_code, ****1234, VISA
API-->>UI: Payment Approved
UI-->>U: "Approved - $45.00"
Note over U, PROC: Refund Flow (Using Token)
U->>UI: Process Refund
UI->>API: POST /refunds/create
Note right of API: {order_id, amount, reason}
API->>API: Retrieve Stored Token
API->>PROC: Refund Request with Token
PROC-->>API: Refund Approved
API-->>UI: Refund Complete
1.18.1 Payment Data Storage Rules
payment_data:
# Data your system STORES
stored:
- transaction_id # Your internal ID
- payment_token # Processor token for refunds
- approval_code # Authorization code
- masked_card_number # Last 4 digits only (****1234)
- card_brand # Visa, Mastercard, Amex, Discover
- entry_method # chip, tap, swipe, manual
- terminal_id # Which terminal processed
- timestamp # When processed
- amount # Transaction amount
# Data your system NEVER stores (PCI prohibited)
prohibited:
- full_card_number # 16-digit PAN
- cvv_cvc # Security code
- track_data # Magnetic stripe data
- pin_block # Encrypted PIN
- emv_data # Chip cryptogram (raw)
1.18.2 Terminal Communication
terminal_integration:
protocol: "cloud_api" # Terminal vendor's cloud service
# Timeout settings
payment_timeout_seconds: 60
connection_timeout_seconds: 10
# Terminal states
states:
- IDLE: "Ready for transaction"
- WAITING_FOR_CARD: "Display amount, await tap/insert"
- PROCESSING: "Communicating with processor"
- APPROVED: "Transaction successful"
- DECLINED: "Transaction declined"
- ERROR: "Communication or hardware error"
- CANCELLED: "Customer or staff cancelled"
# Error handling
on_timeout: "prompt_retry_or_cancel"
on_decline: "display_reason_allow_retry"
on_error: "log_and_alert_manager"
# Void window (same-day before batch)
same_day_void: true
batch_close_time: "23:00" # Auto-batch at 11 PM
1.18.3 Payment Terminal Failure Handling
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant TERM as Payment Terminal
U->>UI: Click "Pay by Card"
UI->>API: POST /payments/initiate
API->>TERM: Send Payment Request
alt Terminal Timeout
TERM--xAPI: No Response (60s)
API-->>UI: "Terminal not responding"
UI-->>U: Options: Retry | Different Terminal | Cash | Cancel
else Terminal Declined
TERM-->>API: Declined (Insufficient Funds)
API-->>UI: "Card Declined: Insufficient Funds"
UI-->>U: Options: Try Another Card | Cash | Cancel
else Terminal Error
TERM-->>API: Error (Hardware Issue)
API-->>UI: "Terminal Error"
UI-->>U: Options: Different Terminal | Cash | Cancel
API->>API: Log Error, Alert Manager
end
1.18.4 Reports: Payment Integration
| Report | Purpose | Key Data Fields |
|---|---|---|
| Payment Terminal Performance | Monitor terminal health | Terminal ID, transaction count, avg response time, error rate, decline rate |
| Decline Rate Report | Track payment failures | Decline reason, frequency, terminal, time of day, retry success rate |
| Batch Settlement Report | Daily batch close summary | Batch date, transaction count, total amount, settlement status |
1.19 Sales User Stories (Epics)
Epic 1.A: Core Sales & Inventory
- Story 1.A.1 (Hybrid Entry): Staff can add items via Scanner (bulk array, max 50 tags) or Barcode (single SKU), with immediate stock validation.
- Story 1.A.2 (Parking): Staff can “Park” a sale (max 5 per terminal, 4-hour TTL) to serve another customer and “Retrieve” it later. Parked items soft-reserve inventory.
- Story 1.A.3 (Mixed Cart): A single transaction can contain both Sales (positive price) and Returns (negative price), calculating a Net Total.
- Story 1.A.4 (Inventory Checks): System validates stock > 0 before adding to cart. Returns trigger
INCREMENTstock event; Sales triggerDECREMENT. - Story 1.A.5 (Price Check Mode): Staff can scan items in “Price Check” mode to display price without adding to cart. Useful for customer inquiries.
Epic 1.B: Pricing & Promotion
- Story 1.B.1 (Smart Promos): The system alerts staff (“Upsell Opportunity”) when a cart is eligible for a promo (e.g., “Buy 2 Get 1 Free”).
- Story 1.B.2 (Granular Discounts): Line-item discounts apply before global discounts. Manager PIN is required for overrides above a certain %.
- Story 1.B.3 (Discount Stacking): System prevents invalid stacking (e.g., cannot use Promo Code on “Clearance” items).
- Story 1.B.4 (Price Tiers): Customers can be assigned price tiers (Retail, Wholesale, VIP, Employee) that apply different base prices before any discounts.
- Story 1.B.5 (Coupon System): System supports both single-use coupons (birthday, referral) and multi-use coupons (promotional codes). Single-use coupons are marked redeemed after use.
- Story 1.B.6 (Discount Order): Discounts apply in strict order: Price Tier → Line Discounts → Auto Promos → Global % → Coupons → Tax → Loyalty Redemptions. Loyalty redemptions apply after tax calculation.
Epic 1.C: Payments & Financials
- Story 1.C.1 (On-Account): Trusted customers can buy on credit. System validates
Current Debt + Pending Layaways + Cart <= Credit Limit. Paying off debt requires Cash/Card (prevents circular credit). - Story 1.C.2 (Layaway): Customers can reserve items with a partial deposit. Inventory is reserved immediately (Status:
RESERVED), but revenue is not fully recognized until completion. - Story 1.C.3 (Split Tender): Transactions support mixed payments (e.g., $20 Cash + Remaining on Card). Customers can use multiple credit cards and combine cash + card(s) in any combination. Each card’s token is stored separately for refund processing.
- Story 1.C.4 (Gift Cards): Staff can sell gift cards, reload existing cards, check balances, and accept gift cards as payment. Partial redemption is supported. Jurisdiction rules apply.
- Story 1.C.5 (Cash Drawer Management): Each shift requires opening float entry, blind cash counts at close, variance tracking, X-report (mid-shift) and Z-report (end-of-shift) generation. Variances outside tolerance require manager approval.
- Story 1.C.6 (Credit Card Payments - SAQ-A): Card payments use semi-integrated terminals. Card data never touches our system. Only tokens, approval codes, and masked card numbers are stored.
- Story 1.C.7 (Payment Failures): When terminal times out, declines, or errors, staff can retry, switch terminals, accept cash, or cancel. All failures are logged.
- Story 1.C.8 (Third-Party Financing - Affirm): Staff can offer Affirm as a payment option. Customer completes financing on their device. Store receives full payment from Affirm immediately. Customer repays Affirm per their loan terms.
- Story 1.C.9 (Multiple Payment Methods): Customers can split payment across multiple credit cards or combine cash + card(s). System tracks each card’s token separately for individual refund processing.
Epic 1.D: Post-Sale & Data
- Story 1.D.1 (Voiding): Voiding is only allowed same business day with drawer open. Voiding reverses inventory, loyalty, and commissions (full reversal). Record is flagged
VOIDED, not deleted. - Story 1.D.2 (Returns): Returns require staff to scan the receipt for system validation before processing. Commission is reversed proportionally based on returned value. Refunds use stored payment tokens or manual terminal processing (staff chooses).
- Story 1.D.3 (Receipts): Staff can choose between Thermal (standard), A4 (invoice), or Gift Receipts (hidden price) at print time.
- Story 1.D.4 (Receipt Reprint): Staff can reprint any historical receipt or email to a different address.
- Story 1.D.5 (History & Export): Managers can filter history by Date/User/Status and export to CSV (limit 1,000 rows for performance).
- Story 1.D.6 (Return Policy Engine): System enforces return and exchange policies that are manually configured in the application’s Settings/Setup module. Policies are per-tenant, per-store, and per-channel (online vs in-store). Example defaults: 30 days full refund with receipt, 31-90 days store credit, no receipt = store credit at current price with manager approval.
- Story 1.D.7 (Dedicated Exchanges): Staff can process exchanges as a single transaction showing item out, item in, and price difference. Exchange records link the original and new items. Commission adjusts for price difference.
Epic 1.E: Special Orders & Transfers
- Story 1.E.1 (Special Orders): Staff can create special orders for out-of-stock items. Customer pays deposit (minimum 50% or configurable). System tracks order status and notifies customer on arrival.
- Story 1.E.2 (Multi-Store Inventory): Staff can view inventory at all store locations (eventually consistent, max 5 min stale). Can request transfers or reserve items at other stores for customer pickup.
- Story 1.E.3 (Transfer/Reserve Payment): Transfers and reservations require customer to pay in full before the request is submitted to the system. This prevents unpaid phantom requests.
- Story 1.E.4 (Hold for Pickup): Fully paid orders can be held for customer pickup with a configurable time limit (default 7 days). System alerts staff when holds expire.
- Story 1.E.5 (Ship to Customer): Staff can ship items from other store locations directly to the customer’s address. System integrates with carrier APIs to calculate real-time shipping costs based on origin store and destination address. Customer pays item price + shipping in full before shipment is initiated. Tracking number is shared with customer via email.
Epic 1.F: Tracking & Commissions
- Story 1.F.1 (Serial Numbers): Designated products require serial number capture at sale. System validates serial hasn’t been previously sold. Serial numbers are searchable for warranty/return lookup.
- Story 1.F.2 (Sales Commissions): Each transaction records the employee ID. Commission amounts are calculated based on sale total and stored for reporting.
- Story 1.F.3 (Commission Reversal): Voided sales reverse commission fully. Returns reverse commission proportionally (returned value / original sale value).
- Story 1.F.4 (Commission Reports): Managers can view commission reports by date range and employee, showing total sales, returns, and net commission earned.
Epic 1.G: Loyalty Programs
- Story 1.G.1 (Points Program): Customers earn points per dollar spent. Points can be redeemed for discounts at configurable rates (e.g., 100 points = $1).
- Story 1.G.2 (Punch Cards): Digital punch cards track qualifying purchases (e.g., 10 coffees = 1 free). Punches are automatically applied based on product categories.
- Story 1.G.3 (Spend Thresholds): Customers earn rewards when spending thresholds are met (e.g., spend $100/month, get $10 off). Rewards can be auto-applied or saved.
- Story 1.G.4 (Loyalty Tiers): Customers automatically upgrade tiers (Bronze → Silver → Gold) based on spend. Higher tiers earn more points or get better rewards.
Epic 1.H: Offline & Resilience
- Story 1.H.1 (Offline Detection): System automatically detects network loss and switches to offline mode with clear visual indicator.
- Story 1.H.2 (Offline Sales): Staff can process sales, returns with receipt, and price checks while offline. Transactions queue locally (max 100).
- Story 1.H.3 (Blocked Offline): System blocks risky operations offline: new customer creation, credit checks, gift card activation/reload, multi-store operations.
- Story 1.H.4 (Auto Sync): When network restores, system automatically syncs queued transactions. Conflicts are flagged for manager review.
- Story 1.H.5 (Conflict Resolution): Manager can review sync conflicts (out of stock, price changes) and approve, modify, or cancel affected transactions.
Epic 1.I: Tax Calculation
- Story 1.I.1 (Jurisdiction Taxes): System calculates tax based on store physical address, applying state + county + city + district rates as applicable.
- Story 1.I.2 (Tax Hierarchy): Product-level tax overrides take priority over customer exemptions, which take priority over default store rates.
- Story 1.I.3 (Tax Exemptions): Customers with valid exemption certificates (resale, non-profit, diplomatic) can be flagged for tax-exempt purchases.
- Story 1.I.4 (Tax Display): Receipt shows tax breakdown by jurisdiction (State Tax: $X, Local Tax: $Y).
1.20 Sales Acceptance Criteria (Gherkin)
Feature: Gift Card Operations
Feature: Gift Card Management
As a retail staff member
I need to sell, reload, and redeem gift cards
So that customers can purchase and use store credit
Background:
Given I am logged into the POS system
And the cash drawer is open
Scenario: Sell a new gift card
Given I have a physical gift card with number "GC-000000001234"
And the gift card status is "INACTIVE"
When I scan the gift card
And I enter load amount "$50.00"
And the customer pays "$50.00"
Then the gift card status should be "ACTIVE"
And the gift card balance should be "$50.00"
And a receipt should print showing "Gift Card Activated: $50.00"
Scenario: Check gift card balance
Given a gift card "GC-000000001234" with balance "$35.50"
When I click "Gift Card Balance"
And I scan the gift card
Then the display should show "Balance: $35.50"
And the display should show expiry info based on jurisdiction
Scenario: Partial redemption
Given a gift card "GC-000000001234" with balance "$50.00"
And a cart total of "$30.00"
When I apply the gift card as payment
Then the gift card balance should become "$20.00"
And the order should be marked "PAID"
And the gift card status should remain "ACTIVE"
Scenario: Full redemption depletes card
Given a gift card "GC-000000001234" with balance "$25.00"
And a cart total of "$25.00"
When I apply the gift card as payment
Then the gift card balance should become "$0.00"
And the gift card status should be "DEPLETED"
Scenario: Insufficient balance prompts partial
Given a gift card "GC-000000001234" with balance "$20.00"
And a cart total of "$50.00"
When I apply the gift card as payment
Then the system should display "Card Balance: $20.00. Apply partial?"
When I confirm partial application
Then "$20.00" should be applied from the gift card
And remaining balance should show "$30.00"
Scenario: Reload existing gift card
Given a gift card "GC-000000001234" with balance "$10.00"
When I scan the gift card
And I select "Reload"
And I enter reload amount "$25.00"
And the customer pays "$25.00"
Then the gift card balance should be "$35.00"
Scenario: Cash out in California
Given the store is located in California
And a gift card "GC-000000001234" with balance "$8.50"
When I scan the gift card
Then the system should show "Eligible for Cash Out: $8.50"
When I process cash out
Then the gift card balance should be "$0.00"
And "$8.50" cash should be given to customer
Feature: Exchange Transactions
Feature: Dedicated Exchange Flow
As a retail staff member
I need to process exchanges efficiently
So that customers can swap items in a single transaction
Background:
Given I am logged into the POS system
And order "ORD-001" exists with item "Blue Shirt" at "$40.00"
And the original sale had commission "$2.00"
Scenario: Even exchange - same price items
Given I load order "ORD-001" for exchange
When I select "Blue Shirt" to exchange OUT
And I scan "Red Shirt" priced at "$40.00" to exchange IN
Then the price difference should show "$0.00"
And I should see "Even Exchange - No Payment Required"
When I complete the exchange
Then a new exchange record should link "ORD-001" to the new order
And inventory for "Blue Shirt" should INCREMENT by 1
And inventory for "Red Shirt" should DECREMENT by 1
And commission should remain unchanged
Scenario: Customer owes money - upgrade
Given I load order "ORD-001" for exchange
When I select "Blue Shirt" ($40.00) to exchange OUT
And I scan "Premium Jacket" priced at "$80.00" to exchange IN
Then the price difference should show "Customer Owes: $40.00"
When the customer pays "$40.00"
And I complete the exchange
Then the exchange should be recorded successfully
And additional commission should be recorded for the $40.00 difference
Scenario: Store owes refund - downgrade
Given I load order "ORD-001" for exchange
When I select "Blue Shirt" ($40.00) to exchange OUT
And I scan "Basic Tee" priced at "$25.00" to exchange IN
Then the price difference should show "Refund to Customer: $15.00"
When I process the refund
And I complete the exchange
Then "$15.00" should be refunded to original payment method
And commission should be reduced proportionally
Feature: Multi-Store Transfer (Full Payment Required)
Feature: Multi-Store Inventory Transfer
As a retail staff member
I need to request transfers from other stores
So that customers can get items not available locally
Background:
Given I am logged into POS at "Store A"
And "Store B" has 5 units of "SKU-12345"
And "Store A" has 0 units of "SKU-12345"
Scenario: Transfer request requires full payment
Given a customer wants "SKU-12345" priced at "$75.00"
When I search for "SKU-12345"
And I click "Check Other Stores"
Then I should see "Store B: 5 units"
And I should see "Last updated: X minutes ago"
When I click "Request Transfer" from "Store B"
Then I should see "Customer must pay in full to process"
When the customer pays "$75.00"
Then a transfer record should be created with status "PAID"
And "Store B" should receive a transfer notification
And I should print a transfer receipt for the customer
Scenario: Transfer blocked without payment
Given a customer wants "SKU-12345" priced at "$75.00"
When I click "Request Transfer" from "Store B"
And the customer declines to pay
Then no transfer record should be created
And "Store B" inventory should remain unchanged
Scenario: Reservation at other store requires full payment
Given a customer wants to pick up "SKU-12345" at "Store B"
When I click "Reserve at Store B"
Then I should see "Customer must pay in full to reserve"
When the customer pays "$75.00"
Then a reservation should be created at "Store B"
And the customer should receive a pickup voucher
And the reservation should expire in 7 days
Feature: Return Policy Engine
Feature: Return Policy Enforcement
As a retail staff member
I need the system to enforce return policies
So that returns are handled consistently
Scenario: Full refund within 30 days with receipt
Given order "ORD-001" was placed 15 days ago
And the customer has the receipt
When I scan the receipt
And I select items to return
Then the system should show "Full Refund Eligible"
And the refund should go to the original payment method
And commission should be reversed proportionally
Scenario: Store credit for 31-90 days
Given order "ORD-001" was placed 45 days ago
And the customer has the receipt
When I scan the receipt
And I select items to return
Then the system should show "Store Credit Only"
And I should issue store credit for the return amount
Scenario: Manager override required beyond 90 days
Given order "ORD-001" was placed 120 days ago
And the customer has the receipt
When I scan the receipt
And I select items to return
Then the system should show "Manager Approval Required"
When a manager enters their PIN
And selects exception reason "Customer Goodwill"
Then the return should proceed
Scenario: No receipt - store credit at current price
Given a customer wants to return "Blue Shirt"
And the customer has no receipt
And "Blue Shirt" current price is "$35.00"
When I scan the item barcode
Then the system should show "No Receipt - Store Credit at Current Price"
And the system should show "Manager Approval Required"
When a manager approves
Then store credit for "$35.00" should be issued
Scenario: Final sale items blocked
Given order "ORD-001" contains item "Clearance Item" marked as "Final Sale"
When I attempt to return "Clearance Item"
Then the system should show "BLOCKED: Final Sale - No Returns"
And the return should not proceed
Scenario: Restocking fee for opened electronics
Given order "ORD-001" contains "Headphones" in category "Electronics"
And the item has been opened
When I process the return
Then the system should show "Restocking Fee: 15%"
And the refund should be reduced by 15%
Feature: Void vs Return Distinction
Feature: Void vs Return Rules
As a retail staff member
I need clear rules for when to void vs return
So that corrections are handled properly
Scenario: Void allowed same day with drawer open
Given order "ORD-001" was completed today
And the cash drawer is still open
When I select order "ORD-001"
And I click "Void"
Then the void should be allowed
And inventory should be reversed immediately
And commission should be fully reversed
And the order status should be "VOIDED"
Scenario: Void blocked after drawer close
Given order "ORD-001" was completed today
And the cash drawer has been closed
When I select order "ORD-001"
And I click "Void"
Then the system should show "Cannot void - drawer closed. Use Return instead."
Scenario: Void blocked next day
Given order "ORD-001" was completed yesterday
When I select order "ORD-001"
And I click "Void"
Then the system should show "Cannot void - different business day. Use Return instead."
Scenario: Return uses proportional commission reversal
Given order "ORD-001" has 3 items totaling "$120.00"
And commission was "$6.00" (5%)
When I return 1 item worth "$40.00"
Then commission should be reduced by "$2.00" (40/120 × $6)
And net commission should be "$4.00"
Feature: Special Orders
Feature: Special Order Management
As a retail staff member
I need to create special orders for out-of-stock items
So that customers can order items not currently available
Scenario: Create special order with deposit
Given "SKU-99999" is out of stock
And it is available for special order at "$100.00"
When I create a special order for customer "John Doe"
And I enter quantity "1"
Then the system should calculate deposit as "$50.00" (50%)
When the customer pays "$50.00" deposit
Then special order "SO-12345" should be created
And the status should be "DEPOSIT_PAID"
And the purchasing team should be notified
Scenario: Complete special order on arrival
Given special order "SO-12345" exists with deposit "$50.00"
And the remaining balance is "$50.00"
When the item arrives and status becomes "READY_FOR_PICKUP"
Then customer "John Doe" should receive a notification
When the customer arrives and pays "$50.00"
Then the special order status should be "COMPLETED"
Feature: Cash Drawer Management
Feature: Cash Drawer Operations
As a retail staff member
I need to manage the cash drawer
So that cash is tracked accurately
Scenario: Open drawer with float
Given the cash drawer is closed
When a manager opens the drawer
And enters opening float "$200.00"
Then the drawer session should start
And the drawer status should be "OPEN"
Scenario: Close drawer with balanced count
Given the drawer is open with float "$200.00"
And cash sales today total "$350.00"
And cash refunds today total "$50.00"
When I click "Close Drawer"
And I perform blind count entering "$500.00"
Then expected amount should be "$500.00"
And variance should be "$0.00"
And the system should show "Drawer Balanced"
And a Z-report should print
Scenario: Variance requires manager approval
Given the drawer is open with float "$200.00"
And expected cash is "$500.00"
When I perform blind count entering "$493.00"
Then variance should be "-$7.00"
And the system should show "Variance: -$7.00 - Manager Approval Required"
When a manager enters PIN and reason "Counting Error"
Then the drawer should close
And the variance should be logged
Scenario: Variance within tolerance auto-approves
Given variance tolerance is set to "$5.00"
And expected cash is "$500.00"
When I perform blind count entering "$497.00"
Then variance should be "-$3.00"
And the system should show "Drawer Balanced" (within tolerance)
Scenario: Run X-Report mid-shift
Given the drawer is open with float "$200.00"
And cash sales so far total "$150.00"
And cash refunds so far total "$20.00"
When I click "X-Report"
Then the X-Report should show opening float "$200.00"
And the X-Report should show cash sales "$150.00"
And the X-Report should show cash refunds "$20.00"
And the X-Report should show expected cash "$330.00"
And the drawer should remain OPEN
And I should be able to continue processing transactions
Scenario: Run multiple X-Reports in one shift
Given the drawer is open
When I run an X-Report at 10:00 AM
And I process more sales
And I run another X-Report at 2:00 PM
Then both X-Reports should complete successfully
And the second X-Report should reflect updated totals
And the drawer should remain OPEN
Feature: Coupon System
Feature: Coupon Redemption
As a retail staff member
I need to apply coupons to transactions
So that customers receive their discounts
Scenario: Apply single-use birthday coupon
Given coupon "BDAY-JOHN-2026" exists
And it is a single-use coupon for "$10 off"
And it has not been redeemed
When I scan the coupon
Then "$10.00" discount should apply
When I complete the sale
Then the coupon status should be "REDEEMED"
Scenario: Reject already-used single-use coupon
Given coupon "BDAY-JOHN-2026" has been redeemed
When I scan the coupon
Then the system should show "Coupon Already Redeemed"
And no discount should apply
Scenario: Apply multi-use promotional coupon
Given coupon "SAVE10" exists
And it is a multi-use coupon for "10% off"
And it has been used 50 times (limit 1000)
When I scan the coupon
Then "10% off" discount should apply
When I complete the sale
Then the coupon usage count should be 51
Scenario: Reject expired coupon
Given coupon "SUMMER2025" expired on "2025-08-31"
When I scan the coupon
Then the system should show "Coupon Expired"
And no discount should apply
Feature: Loyalty Programs
Feature: Flexible Loyalty Programs
As a retail staff member
I need to manage customer loyalty
So that customers are rewarded for purchases
Scenario: Earn points per dollar
Given customer "Jane Doe" is attached to the sale
And the loyalty program awards 1 point per $1
And the cart total is "$45.00"
When I complete the sale
Then "Jane Doe" should earn 45 points
And the receipt should show "Points Earned: 45"
Scenario: Redeem points for discount
Given customer "Jane Doe" has 500 points
And redemption rate is 100 points = $1
When I click "Redeem Points"
And I redeem 500 points
Then "$5.00" discount should apply
And "Jane Doe" points should become 0
Scenario: Punch card completion
Given customer "Jane Doe" has a coffee punch card
And she has 9 of 10 punches
And the cart contains 1 coffee
When I complete the sale
Then the punch card should show "FREE ITEM EARNED!"
And the 10th coffee should be free
And a new punch card should start
Scenario: Tier upgrade on spend threshold
Given customer "Jane Doe" is "BRONZE" tier
And she has spent "$950" this year
And the cart total is "$100"
When I complete the sale
Then "Jane Doe" should be upgraded to "SILVER"
And the system should show "Customer upgraded to Silver!"
And she should now earn 1.5x points
Feature: Offline Operations
Feature: Offline Mode Operations
As a retail staff member
I need to continue serving customers when network is down
So that business is not interrupted
Background:
Given I am logged into the POS system
And the cash drawer is open
Scenario: Detect offline and show indicator
Given the network connection is lost
Then the UI should show "OFFLINE MODE" indicator
And the indicator should be prominently visible
Scenario: Process sale while offline
Given I am in offline mode
When I scan items and complete a sale
Then the transaction should be queued locally
And the receipt should print with "OFFLINE" watermark
And I should see "1 transaction pending sync"
Scenario: Block risky operations offline
Given I am in offline mode
When I try to create a new customer
Then the system should show "Not available offline"
When I try to activate a gift card
Then the system should show "Not available offline"
When I try to check multi-store inventory
Then the system should show "Not available offline"
Scenario: Auto-sync when network restores
Given I am in offline mode
And I have 3 queued transactions
When the network connection is restored
Then the UI should show "SYNCING..."
And transactions should sync automatically
When sync completes
Then the UI should show "All transactions synced"
Scenario: Handle sync conflict
Given I am in offline mode
And I sold the last unit of "SKU-123"
When the network restores
And another store also sold the last unit
Then the system should show "Conflict: SKU-123 out of stock"
And I should be prompted for manager review
When the manager approves backorder
Then the transaction should complete with backorder status
Feature: Payment Integration
Feature: SAQ-A Payment Processing
As a retail staff member
I need to process card payments securely
So that customer card data is protected
Scenario: Successful card payment
Given a cart total of "$45.00"
When I click "Pay by Card"
Then the terminal should display "$45.00"
When the customer taps their card
And the payment is approved
Then I should see "Approved - $45.00"
And the receipt should show "VISA ****1234"
And no full card number should be stored
Scenario: Card declined - insufficient funds
Given a cart total of "$200.00"
When I click "Pay by Card"
And the customer's card is declined for insufficient funds
Then I should see "Card Declined: Insufficient Funds"
And I should have options: "Try Another Card" | "Cash" | "Cancel"
Scenario: Terminal timeout
Given a cart total of "$45.00"
When I click "Pay by Card"
And the terminal does not respond within 60 seconds
Then I should see "Terminal not responding"
And I should have options: "Retry" | "Different Terminal" | "Cash" | "Cancel"
Scenario: Refund using stored token
Given order "ORD-001" was paid by card
And the payment token is stored
When I process a refund for "ORD-001"
Then the refund should use the stored token
And I should not need to re-enter card details
And the receipt should show "Refund to VISA ****1234"
Scenario: Pay with Affirm financing
Given a cart total of "$500.00"
When I click "Pay with Affirm"
Then a QR code or redirect should display for the customer
When the customer completes Affirm application on their device
And Affirm approves the loan
Then I should see "Payment Approved (Affirm)"
And the full $500.00 should be received from Affirm
And no card data should be stored
And an Affirm charge_id should be stored
Scenario: Split payment across two credit cards
Given a cart total of "$200.00"
When I click "Pay by Card"
And the customer pays "$100.00" with first card
Then remaining balance should show "$100.00"
When I click "Pay by Card" again
And the customer pays "$100.00" with second card
Then the order should be marked "PAID"
And two separate payment tokens should be stored
Scenario: Combine cash and card payment
Given a cart total of "$150.00"
When I accept "$50.00" cash
Then remaining balance should show "$100.00"
When the customer pays "$100.00" by card
Then the order should be marked "PAID"
Feature: Tax Calculation
Feature: Tax Calculation Engine
As a retail staff member
I need accurate tax calculation
So that the correct tax is collected
Scenario: Standard Virginia tax
Given the store is located in Richmond, Virginia
And the cart contains taxable items totaling "$100.00"
When I calculate tax
Then state tax should be "$4.30" (4.3%)
And local tax should be "$1.00" (1.0%)
And total tax should be "$5.30"
Scenario: Northern Virginia regional tax
Given the store is located in Fairfax, Virginia
And the cart contains taxable items totaling "$100.00"
When I calculate tax
Then state tax should be "$4.30" (4.3%)
And local tax should be "$1.00" (1.0%)
And regional tax should be "$0.70" (0.7%)
And total tax should be "$6.00"
Scenario: Product-level tax override
Given the cart contains "Grocery Item" in tax category "grocery_food"
And "Grocery Item" price is "$20.00"
When I calculate tax
Then tax on "Grocery Item" should be "$0.30" (1.5% reduced rate)
Scenario: Customer tax exemption
Given customer "ABC Nonprofit" has tax exemption status "NONPROFIT"
And the cart contains taxable items totaling "$100.00"
When I attach customer "ABC Nonprofit" to the sale
And I calculate tax
Then total tax should be "$0.00"
And the receipt should show "Tax Exempt: 501(c)(3)"
Scenario: Tax hierarchy - product override takes priority
Given customer "ABC Nonprofit" has tax exemption status
And the cart contains "Prepared Food" (10% tax rate)
When I calculate tax
Then the product-level 10% rate should apply
Because product-level overrides take priority over customer exemption
2. Customers Module
2.1 Customer Management Workflow
Scope: Creating Profiles, Merging Duplicates, Tax Logic, Groups, and Deletion Guards.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Phase 1: Creation & Maintenance
U->>UI: Search Customer (Name / Phone / Email)
UI->>API: GET /customers/search
alt Customer Found
API-->>UI: Return Profile
UI-->>U: Display Customer Details + Group + Price Tier
else Create New
U->>UI: Enter Details (Name, Phone, Email)
U->>UI: Enter Physical Address (Shipping)
U->>UI: Enter Billing Address (if different)
opt Customer Group Assignment
U->>UI: Select Group (Retail/Wholesale/VIP/Staff)
Note right of UI: Group determines Price Tier
end
opt Tax Assignment
U->>UI: Select Custom Tax Rate (e.g., "Tax Exempt")
Note right of UI: Overrides Default Store Tax
end
opt Communication Preferences
U->>UI: Set Email Opt-In (Y/N)
U->>UI: Set SMS Opt-In (Y/N)
U->>UI: Set Preferred Contact Method
end
opt Customer Notes
U->>UI: Enter Size Preferences
U->>UI: Enter Free-Form Notes
end
UI->>API: POST /customers/create
API->>DB: Insert Record
end
2.2 Customer Groups & Tiers
Scope: Automatic and manual group assignment with price tier implications.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Customer Group Management
alt Manual Group Assignment
U->>UI: Open Customer Profile
U->>UI: Click "Change Group"
U->>UI: Select Group (Retail/Wholesale/VIP/Staff)
UI->>API: PATCH /customers/{id}/group
API->>DB: Update Customer Group
API->>DB: Update Price Tier (Based on Group)
API-->>UI: Group Updated
end
Note over API, DB: Automatic Tier Upgrades (Background Job)
API->>DB: Check Customer Spend Totals
alt Spend >= Gold Threshold ($5,000/year)
API->>DB: Upgrade to Gold Tier
API-->>UI: Notify: "Customer upgraded to Gold!"
else Spend >= Silver Threshold ($1,000/year)
API->>DB: Upgrade to Silver Tier
end
Note right of DB: Tier Benefits:
Note right of DB: Bronze: 1x points
Note right of DB: Silver: 1.5x points + 5% discount
Note right of DB: Gold: 2x points + 10% discount
2.3 Customer Notes & Preferences
Scope: Structured fields and free-form notes for customer preferences.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
U->>UI: Open Customer Profile
U->>UI: Click "Notes & Preferences"
Note over UI, DB: Structured Preferences
U->>UI: Enter Clothing Size (S/M/L/XL)
U->>UI: Enter Shoe Size
U->>UI: Select Color Preferences
U->>UI: Enter Brand Preferences
Note over UI, DB: Free-Form Notes
U->>UI: Enter Notes (e.g., "Prefers classic styles, allergic to wool")
UI->>API: PATCH /customers/{id}/preferences
API->>DB: Update Preference Fields
API-->>UI: Saved
Note over U, UI: Notes Display at POS
U->>UI: Attach Customer to Sale
UI->>API: GET /customers/{id}
API-->>UI: Return Profile with Notes
UI-->>U: Display: "Notes: Prefers classic styles, Size M"
2.4 Communication Preferences
Scope: Marketing consent, contact preferences, and opt-out management.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Communication Preference Management
U->>UI: Open Customer Profile -> Communications
UI-->>U: Display Current Settings:
Note right of UI: Email Marketing: [ON/OFF]
Note right of UI: SMS Marketing: [ON/OFF]
Note right of UI: Preferred Contact: [Email/Phone/SMS]
Note right of UI: Do Not Contact: [Y/N]
U->>UI: Toggle Email Marketing ON
U->>UI: Toggle SMS Marketing OFF
U->>UI: Set Preferred: Email
UI->>API: PATCH /customers/{id}/communication-prefs
API->>DB: Update Communication Preferences
API->>DB: Log Consent Change (Audit Trail)
API-->>UI: Preferences Saved
Note over API, DB: Privacy Compliance
Note right of DB: All consent changes logged with timestamp
Note right of DB: Customer can request full data export
Note right of DB: "Do Not Contact" blocks all outreach
2.5 Advanced Customer Actions
Scope: Merge duplicates, safe deletion, data export, and privacy compliance.
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Phase 2: Advanced Actions (Merge & Delete)
alt Action: Merge Duplicates
U->>UI: Select "Source" (Duplicate) & "Target" (Primary)
U->>UI: Click "Merge Customers"
UI->>API: POST /customers/merge
par Data Transfer
API->>DB: Move History, Loyalty, Balance to Target
API->>DB: Merge Notes (Append Source to Target)
API->>DB: Keep Higher Tier
API->>DB: Soft-Delete "Source" Profile
end
API-->>UI: Merge Success
end
alt Action: Delete Customer
U->>UI: Click "Delete Customer"
UI->>API: GET /customers/{id}/balance-check
alt Has Debt or Open Layaway
API-->>UI: Error: "Cannot Delete - Outstanding Balance"
UI-->>U: Alert: "Settle Balance First"
else Safe to Delete
UI->>API: DELETE /customers/{id}
API->>DB: Anonymize Personal Data
API->>DB: Retain Sales History (Linked to "Anonymous")
API-->>UI: Deletion Success
end
end
alt Action: Export Data
U->>UI: Filter List -> Click "Export CSV"
UI->>API: POST /customers/export
Note right of UI: Limit 1000 rows
API-->>UI: Download CSV File
end
alt Action: Data Subject Request (Privacy)
U->>UI: Click "Privacy Request"
U->>UI: Select Type: Export / Delete / Restrict
UI->>API: POST /customers/{id}/privacy-request
API->>DB: Log Request with Timestamp
API-->>UI: Request ID Generated
Note right of API: Must complete within 30 days
end
2.6 Customer Self-Service
Scope: Customer-facing loyalty balance and preference management.
sequenceDiagram
autonumber
participant C as Customer
participant UI as Self-Service Kiosk / App
participant API as Backend
participant DB as DB
Note over C, DB: Customer Self-Service Options
alt Check Loyalty Balance
C->>UI: Enter Phone Number or Scan Card
UI->>API: GET /customers/lookup?phone={phone}
API->>DB: Find Customer
API-->>UI: Return Loyalty Summary
UI-->>C: Display: "Points: 450 | Tier: Silver | $4.50 available"
end
alt Update Preferences
C->>UI: Login (Phone + PIN)
UI->>API: GET /customers/{id}/preferences
API-->>UI: Return Current Preferences
C->>UI: Update Email / SMS Opt-In
UI->>API: PATCH /customers/{id}/communication-prefs
API->>DB: Update & Log Consent Change
API-->>UI: Saved
end
alt Request Data Export
C->>UI: Click "Download My Data"
UI->>API: POST /customers/{id}/data-export
API->>DB: Queue Export Job
API-->>UI: "Export will be emailed within 24 hours"
end
2.6.1 Reports: Customer Module
| Report | Purpose | Key Data Fields |
|---|---|---|
| Customer Activity Report | Track customer engagement | Customer, last purchase date, total spend, visit frequency, loyalty tier |
| New Customer Report | Monitor customer acquisition | New customers by period, acquisition source, first purchase value |
| Customer Merge Audit Log | Track merge operations | Source ID, target ID, merge date, merged by, data transferred |
| Customer Group Distribution | Breakdown by group/tier | Group, customer count, avg spend, revenue contribution |
| Inactive Customer Report | Identify disengaged customers | Customer, last activity, days inactive, lifetime value, tier |
Email Template: TMPL-WELCOME
| Field | Value |
|---|---|
| Trigger | New customer profile created |
| Recipient | Customer (if email provided and opt-in) |
| Content | Welcome message, loyalty program details, store locations |
Email Template: TMPL-TIER-UPGRADE
| Field | Value |
|---|---|
| Trigger | Customer tier upgraded (e.g., Bronze → Silver) |
| Recipient | Customer |
| Content | New tier name, benefits unlocked, points multiplier, discount percentage |
2.7 Customer User Stories (Epics)
Epic 2.A: Profile & Data Management
- Story 2.A.1 (Detailed Profile): Staff can store distinct “Shipping” (Physical) and “Billing” addresses for a customer to support delivery and invoicing.
- Story 2.A.2 (Duplicate Handling / Merge): Managers can merge two customer profiles. The system must transfer all Sales History, Loyalty Points, Account Balances, and Notes to the “Primary” profile and archive the “Duplicate”.
- Story 2.A.3 (Safe Deletion): The system must block deletion if the customer has an active “On Account” debt or an open “Layaway”. If deleted, their historical sales must remain but be anonymized.
- Story 2.A.4 (Export): Staff can export the customer list to CSV for marketing, limited to 1,000 rows per batch for system stability.
- Story 2.A.5 (Customer Notes): Staff can record structured preferences (size, color, brand) and free-form notes on customer profiles. Notes are displayed when customer is attached to a sale.
Epic 2.B: Financial & Tax Settings
- Story 2.B.1 (Custom Tax Rates): Managers can assign a specific tax exemption status (e.g., “Reseller”, “Non-Profit”) to a customer profile. This status overrides the Store Default Tax when the hierarchy allows.
- Story 2.B.2 (Credit Limits): Managers can set a hard “Credit Limit”. The POS must block any “On Account” sale that pushes (Current Debt + Pending Layaways + Cart) over this limit.
- Story 2.B.3 (Loyalty Adjustments): Managers can manually adjust loyalty points (e.g., +50 “Sorry for wait”). A mandatory reason note is required for audit trails.
Epic 2.C: Customer Groups & Tiers
- Story 2.C.1 (Customer Groups): Customers can be assigned to groups (Retail, Wholesale, VIP, Staff). Each group maps to a price tier that determines base pricing.
- Story 2.C.2 (Automatic Tier Upgrades): Customers automatically upgrade tiers (Bronze → Silver → Gold) when spend thresholds are met. Upgrades can also be manually assigned by managers.
- Story 2.C.3 (Tier Benefits): Each tier has configurable benefits: point multipliers, automatic discounts, and early access to sales.
- Story 2.C.4 (Price Tiers): Different customer groups see different base prices (not just discounts). Wholesale customers may see cost+markup pricing while retail sees standard pricing.
Epic 2.D: Communication & Preferences
- Story 2.D.1 (Communication Consent): Customers can opt-in or opt-out of email and SMS marketing. All consent changes are logged for compliance.
- Story 2.D.2 (Preferred Contact): Customers can specify their preferred contact method (Email, Phone, SMS). Staff can see this preference when reaching out.
- Story 2.D.3 (Do Not Contact): Customers can be flagged as “Do Not Contact” which blocks all marketing outreach while preserving transaction notifications.
Epic 2.E: Privacy & Compliance
- Story 2.E.1 (Data Export): Customers can request a full export of their personal data. Export must be provided within 30 days.
- Story 2.E.2 (Right to Deletion): Customers can request deletion of their personal data. System anonymizes records while preserving transaction history for accounting.
- Story 2.E.3 (Consent Audit Trail): All marketing consent changes are logged with timestamp and source for regulatory compliance.
- Story 2.E.4 (Data Retention): Customer data is retained according to configurable retention policies. Inactive customers can be auto-anonymized after retention period.
Epic 2.F: Self-Service
- Story 2.F.1 (Loyalty Balance Check): Customers can check their loyalty points balance via kiosk, app, or website using phone number lookup.
- Story 2.F.2 (Preference Update): Customers can update their communication preferences (opt-in/opt-out) via self-service channels.
- Story 2.F.3 (Data Request): Customers can submit data export or deletion requests via self-service, which are queued for staff processing.
2.8 Customer Acceptance Criteria (Gherkin)
Feature: Customer Groups and Tiers
Feature: Customer Group Management
As a retail manager
I need to assign customers to groups
So that they receive appropriate pricing and benefits
Scenario: Assign customer to wholesale group
Given customer "ABC Corp" exists
When I open the customer profile
And I click "Change Group"
And I select "Wholesale"
Then the customer group should be "Wholesale"
And the price tier should update to "Wholesale Pricing"
Scenario: Automatic tier upgrade on spend
Given customer "Jane Doe" is "BRONZE" tier
And she has spent "$4,900" this year
And annual spend threshold for Silver is "$1,000"
And annual spend threshold for Gold is "$5,000"
When she makes a purchase of "$150"
Then her annual spend becomes "$5,050"
And she should be upgraded to "GOLD" tier
And she should now earn 2x points
And she should receive 10% automatic discount
Feature: Customer Merge
Feature: Merge Duplicate Customers
As a retail manager
I need to merge duplicate customer profiles
So that customer data is consolidated
Scenario: Merge two customer profiles
Given customer "John Doe" (ID: 100) has:
| Loyalty Points | 500 |
| Account Balance | $50.00 |
| Tier | Silver |
And customer "J. Doe" (ID: 101) has:
| Loyalty Points | 200 |
| Account Balance | $25.00 |
| Tier | Bronze |
When I select "J. Doe" as source and "John Doe" as target
And I click "Merge Customers"
Then "John Doe" should have:
| Loyalty Points | 700 |
| Account Balance | $75.00 |
| Tier | Silver |
And "J. Doe" profile should be archived
And sales history from both should be under "John Doe"
Feature: Safe Customer Deletion
Feature: Customer Deletion Guards
As a retail manager
I need deletion to be blocked when unsafe
So that we don't lose important data
Scenario: Block deletion with outstanding debt
Given customer "John Doe" has account balance "$150.00"
When I click "Delete Customer"
Then the system should show "Cannot Delete - Outstanding Balance"
And the deletion should be blocked
Scenario: Block deletion with open layaway
Given customer "John Doe" has an active layaway order
When I click "Delete Customer"
Then the system should show "Cannot Delete - Open Layaway"
And the deletion should be blocked
Scenario: Safe deletion anonymizes data
Given customer "John Doe" has no debt or layaway
And "John Doe" has 5 historical orders
When I click "Delete Customer"
And I confirm the deletion
Then personal data should be anonymized
And the 5 orders should remain linked to "Anonymous Customer"
Feature: Communication Preferences
Feature: Communication Preference Management
As a retail staff member
I need to manage customer communication preferences
So that we comply with privacy regulations
Scenario: Update email opt-in
Given customer "Jane Doe" has email marketing OFF
When I open communications preferences
And I toggle email marketing ON
Then email marketing should be ON
And a consent change should be logged with timestamp
Scenario: Do Not Contact blocks outreach
Given customer "John Doe" is flagged "Do Not Contact"
When the marketing system attempts to send email
Then the email should be blocked
And no marketing should be sent
Scenario: Transaction notifications still sent
Given customer "John Doe" is flagged "Do Not Contact"
When he makes a purchase
Then a receipt email should still be sent
Because transaction notifications are not marketing
Feature: Privacy Compliance
Feature: Customer Privacy Rights
As a customer
I need to exercise my privacy rights
So that my data is protected
Scenario: Request data export
Given I am customer "Jane Doe"
When I request a data export
Then a request should be logged
And I should receive confirmation
And my data should be emailed within 30 days
Scenario: Request data deletion
Given I am customer "Jane Doe"
And I have no outstanding balance or layaway
When I request data deletion
Then my personal data should be anonymized
And my transaction history should be preserved (anonymized)
And I should receive confirmation
Scenario: Deletion blocked with balance
Given I am customer "Jane Doe"
And I have account balance "$50.00"
When I request data deletion
Then the system should show "Please settle outstanding balance first"
And my data should not be deleted
Feature: Customer Self-Service
Feature: Customer Self-Service
As a customer
I want to check my loyalty status
So that I know my rewards
Scenario: Check loyalty balance
Given I am customer with phone "555-1234"
And I have 450 loyalty points
And I am Silver tier
When I enter my phone number at the kiosk
Then I should see "Points: 450"
And I should see "Tier: Silver"
And I should see "$4.50 available for redemption"
Scenario: Update marketing preferences
Given I am logged into self-service
And email marketing is ON
When I toggle email marketing OFF
Then email marketing should be OFF
And a consent change should be logged
3. Catalog Module
3.1 Product Types & Data Model
Scope: Core POS Catalog – all product entries, types, data fields, and the relationships between them. Every item sold, bundled, or serviced flows through the catalog. The data model supports four distinct product types that cover the full range of retail merchandise: individually tracked goods, multi-dimension variants, composite bundles, and non-inventory services. Beyond the standard field set, the model supports tenant-defined custom attributes, product templates for rapid creation, and a matrix management interface for efficient variant operations.
Cross-Reference: See Module 5, Section 5.10 for UoM configuration and Section 5.12 for custom fields.
3.1.1 Product Types
| Type | Description | Inventory Tracked | Example |
|---|---|---|---|
| Standard | Single product, one SKU, one price | Yes | A branded t-shirt, a candle, a phone case |
| Variant (Parent + Children) | Parent product with up to 3 variant dimensions (e.g., Size, Color, Material). Each combination creates a child SKU with independent inventory and optional price overrides | Yes (per child) | “Classic Oxford Shirt” parent with children: S/Blue, M/Blue, L/White, etc. |
| Composite / Bundle | A kit of component products sold together. Bundle price is set independently – it does NOT need to equal the sum of component prices | Yes (per component, decremented on sale) | “Gift Set” = Candle + Soap + Box for $39.99 (components total $48 individually) |
| Service | Non-inventory item representing labor or a service. No stock tracking, no physical attributes | No | Alterations, gift wrapping, custom engraving, hemming |
3.1.2 Product Class Diagram
classDiagram
class Product {
+UUID id
+String sku
+String name
+String product_type
+Decimal base_price
+Decimal cost
+String lifecycle_status
+Boolean shippable
+UUID package_type_id
+Enum selling_uom
+Enum purchasing_uom
+Decimal uom_conversion_factor
+UUID season_id
+UUID brand_id
+DateTime created_at
+DateTime updated_at
}
class StandardProduct {
+String barcode
+String[] alternate_barcodes
+Boolean track_inventory
+Integer low_stock_threshold
}
class VariantParent {
+String[] dimension_names
+Integer dimension_count
+String style_number
+String[] demographics_age_group
+String[] demographics_gender
+String origin
+String fabric
+UUID season_id
}
class VariantChild {
+UUID parent_id
+String dimension_1_value
+String dimension_2_value
+String dimension_3_value
+String barcode
+Decimal price_override
+Decimal msrp
+Boolean track_inventory
+Boolean deletion_protected
}
class CompositeProduct {
+Decimal bundle_price
+Boolean allow_component_substitution
}
class BundleComponent {
+UUID composite_id
+UUID component_product_id
+Integer quantity
+Decimal component_cost_allocation
}
class ServiceProduct {
+Integer estimated_minutes
+Boolean requires_appointment
+String service_category
}
class CustomAttributeDefinition {
+UUID id
+String name
+Enum type
+Boolean required
+UUID tenant_id
}
class ProductCustomAttribute {
+UUID product_id
+UUID definition_id
+String value
}
class PackageType {
+UUID id
+String name
+Decimal length
+Decimal width
+Decimal height
+Decimal empty_weight
+UUID tenant_id
}
Product <|-- StandardProduct
Product <|-- VariantParent
Product <|-- CompositeProduct
Product <|-- ServiceProduct
VariantParent "1" --> "*" VariantChild : has children
CompositeProduct "1" --> "*" BundleComponent : contains
BundleComponent "*" --> "1" Product : references component
Product "0..*" --> "0..*" ProductCustomAttribute : has custom values
ProductCustomAttribute "*" --> "1" CustomAttributeDefinition : defined by
Product "*" --> "0..1" PackageType : ships in
3.1.3 Full Product Data Model
| Group | Field | Type | Required | Description |
|---|---|---|---|---|
| Identity | id | UUID | Yes | Primary key, system-generated |
sku | String | Yes | Unique stock keeping unit per tenant | |
barcode | String | No | Primary barcode (UPC-A, EAN-13, or internal) | |
alternate_barcodes[] | String[] | No | Additional barcodes (vendor SKU, alternate UPC, etc.) | |
| Description | name | String | Yes | Display name (max 255 chars) |
short_description | String | No | Brief summary for POS display (max 500 chars) | |
long_description | String | No | Full marketing description (max 5000 chars) | |
tags[] | String[] | No | Freeform tags for search and filtering | |
category_id | UUID | Yes | Reference to category hierarchy | |
| Pricing | base_price | Decimal(10,2) | Yes | Default selling price |
cost | Decimal(10,2) | No | Cost of goods (landed cost) | |
compare_at_price | Decimal(10,2) | No | Original price for “was/now” display | |
tax_code | String | No | Tax category override (e.g., “grocery_food”, “clothing_exempt”) | |
| Physical | weight | Decimal(8,3) | No | Product weight for shipping calculation |
weight_unit | Enum | No | lb, oz, kg, g | |
length | Decimal(8,2) | No | Package length | |
width | Decimal(8,2) | No | Package width | |
height | Decimal(8,2) | No | Package height | |
dimension_unit | Enum | No | in, cm | |
| Inventory | track_inventory | Boolean | Yes | Whether to track stock levels (default: true) |
allow_negative | Boolean | Yes | Allow sales when stock is zero (default: false) | |
low_stock_threshold | Integer | No | Alert threshold per location (default: 5) | |
| Media | images[] | URL[] | No | Array of image URLs |
primary_image_id | UUID | No | Reference to primary display image | |
| Status | lifecycle_status | Enum | Yes | DRAFT, ACTIVE, DISCONTINUED, ARCHIVED |
| Timestamps | created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp | |
published_at | DateTime | No | When product went Active | |
discontinued_at | DateTime | No | When product was discontinued | |
archived_at | DateTime | No | When product was archived | |
| Audit | created_by | UUID | Yes | User who created the record |
updated_by | UUID | Yes | User who last modified the record | |
| Retail Attributes | style_number | String | No | Manufacturer or internal style identifier (e.g., “NXJ1078”) |
demographics_age_group | String[] | No | Target age groups: Adult, Youth, Kids, Infant, Toddler | |
demographics_gender | String[] | No | Target genders: Men, Women, Unisex, Boys, Girls | |
origin | String | No | Country or region of manufacture: US, Imported, EU, or custom value | |
fabric | String | No | Primary material composition (e.g., “100% Cotton”, “60% Polyester / 40% Cotton”) | |
season_id | UUID | No | FK to Season – assigns product to a buying season (null = year-round core) | |
brand_id | UUID | No | FK to Brand – the brand label on the product (distinct from vendor/supplier) | |
| Shipping | shippable | Boolean | Yes | Whether this product can be shipped (default: false). When true, weight and dimensions become mandatory |
package_type_id | UUID | No | FK to PackageType – pre-defined package dimensions (Box, Envelope, Satchel, etc.) | |
| Unit of Measure | selling_uom | Enum | Yes | Unit customers buy in (default: EACH). Values: EACH, PAIR, SET, YARD, FOOT, METER, LB, KG, OZ, LITER |
purchasing_uom | Enum | No | Unit purchased from vendor – can differ from selling_uom (e.g., buy by CASE, sell by EACH) | |
uom_conversion_factor | Decimal(10,4) | No | Conversion ratio from purchasing_uom to selling_uom (e.g., 12.0000 means 1 case = 12 each) | |
| Variant-Specific (VariantChild) | msrp | Decimal(10,2) | No | Manufacturer’s Suggested Retail Price – used for “Compare at MSRP” display and margin analysis |
Business Rules – Shipping:
- IF
shippable = trueTHENweight,length,width, andheightare all mandatory. The system rejects save attempts where shippable is enabled but physical dimensions are missing. - IF
package_type_idis set, the package dimensions auto-populate from the PackageType record but can be overridden at the product level.
Business Rules – Unit of Measure:
- IF
purchasing_uomdiffers fromselling_uom, thenuom_conversion_factoris mandatory. - Receiving inventory via Purchase Order uses
purchasing_uom; inventory levels and POS transactions useselling_uom. The system automatically converts quantities using the conversion factor.
3.1.4 Custom Attributes
Scope: Tenant-defined key/value custom fields that extend the standard product data model. Custom attributes allow each tenant to capture business-specific data (e.g., “Certification Level”, “Country of Origin Detail”, “Season Year”) without schema changes.
Custom Attribute Definition Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Display label (e.g., “Certification”, “Import Code”) |
type | Enum | Yes | TEXT, NUMBER, LIST, BOOLEAN |
list_values[] | String[] | No | Allowed values when type = LIST (e.g., [“Organic”, “Fair Trade”, “None”]) |
required | Boolean | Yes | Whether this attribute must be filled on every product (default: false) |
searchable | Boolean | Yes | Whether this attribute is indexed for catalog search (default: false) |
filterable | Boolean | Yes | Whether this attribute appears as a filter in the catalog UI (default: false) |
display_order | Integer | Yes | Sort position in the product edit form |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
ProductCustomAttribute Junction Table
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | FK to products table |
definition_id | UUID | Yes | FK to custom_attribute_definitions table |
text_value | String | No | Value when definition type = TEXT |
number_value | Decimal(15,4) | No | Value when definition type = NUMBER |
list_value | String | No | Selected value when definition type = LIST (must match one of list_values[]) |
boolean_value | Boolean | No | Value when definition type = BOOLEAN |
Business Rules:
- Limit: Up to 50 custom attribute definitions per tenant. Attempting to create a 51st returns an error with guidance to archive unused attributes.
- Indexing: All attributes marked
searchable = trueare indexed via a GIN index on a materialized JSONB representation for fast full-text search. - Filtering: Attributes marked
filterable = trueappear as sidebar filters in the Admin Portal product list and can be used in smart collection rules. - Validation: When type = LIST, the system rejects any value not present in
list_values[]. When type = NUMBER, the system validates numeric format. Whenrequired = true, product save is blocked until a value is provided. - Inheritance: Custom attributes are set at the parent product level. Variant children inherit parent custom attributes and cannot override them individually.
3.1.5 Product Templates & Cloning
Scope: Accelerating product creation by cloning existing products and by applying category-based templates with pre-filled default values. Templates reduce repetitive data entry for categories with consistent attributes (e.g., all t-shirts share the same weight range, tax code, and fabric type).
Product Cloning
- Clone any product to create a new DRAFT with a new auto-generated SKU. All fields are copied except identity fields (
id,sku,barcode), timestamps (created_at,updated_at,published_at), and audit fields (created_byis set to the cloning user). - Cloned products always start in
DRAFTstatus regardless of the source product’s status. - For variant products, cloning copies the parent AND all child variants. Each child receives a new auto-generated SKU. Inventory levels are NOT copied (all set to zero).
- Cloned products receive a default name of “{Original Name} (Copy)” which staff must rename before publishing.
Category-Based Templates
Templates save pre-filled default values per category so that new products created in that category start with sensible defaults already populated.
Template Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Template display name (e.g., “Men’s T-Shirt Defaults”) |
category_id | UUID | Yes | FK to product_categories – the category this template applies to |
default_values | JSON | Yes | Key-value map of field names to default values |
tenant_id | UUID | Yes | Owning tenant |
created_by | UUID | Yes | User who created the template |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Example default_values JSON:
{
"fabric": "100% Cotton",
"weight": 0.35,
"weight_unit": "lb",
"tax_code": "clothing_exempt",
"selling_uom": "EACH",
"shippable": true,
"track_inventory": true,
"low_stock_threshold": 5,
"demographics_gender": ["Men"],
"demographics_age_group": ["Adult"],
"origin": "Imported",
"custom_attributes": {
"Certification": "None",
"Care Instructions": "Machine wash cold"
}
}
Business Rules:
- One template per category per tenant. If a template already exists for a category, staff must edit the existing template rather than create a duplicate.
- When a product is created and assigned to a category that has a template, the system auto-fills all fields from
default_values. Staff can override any auto-filled value before saving. - Templates do not retroactively update existing products. They apply only at creation time.
- Template fields that conflict with required product fields are ignored (e.g., a template cannot set
nameorsku).
3.1.6 Product Matrix Management
Scope: A grid-based interface for managing variant products efficiently. The matrix view presents all variant combinations in a spreadsheet-like layout, enabling rapid price/cost/stock editing and bulk variant creation.
Matrix Grid View
For variant products with two dimensions, the matrix displays:
- Rows = Dimension 1 values (e.g., sizes: S, M, L, XL, XXL)
- Columns = Dimension 2 values (e.g., colors: Red, Blue, Navy, Black)
- Each cell shows: price, cost, and current stock level for that combination
| Red | Blue | Navy | Black |
---------------------------------------------------------------------------
S | $29 / $12 | $29 / $12 | $29 / $12 | $29 / $12 |
| Stock: 14 | Stock: 11 | Stock: 9 | Stock: 7 |
---------------------------------------------------------------------------
M | $29 / $12 | $29 / $12 | $29 / $12 | $29 / $12 |
| Stock: 22 | Stock: 18 | Stock: 15 | Stock: 13 |
---------------------------------------------------------------------------
L | $29 / $12 | $29 / $12 | $29 / $12 | $29 / $12 |
| Stock: 19 | Stock: 16 | Stock: 12 | Stock: 10 |
---------------------------------------------------------------------------
XL | $32 / $13 | $32 / $13 | $32 / $13 | $32 / $13 |
| Stock: 8 | Stock: 6 | Stock: 5 | Stock: 4 |
Inline Editing:
- Click any cell to edit price, cost, or stock level directly in the grid.
- Tab key moves to the next cell. Enter confirms the edit.
- Changed cells are highlighted until saved. A single “Save All Changes” action persists all edits in one batch.
Bulk-Create Variants:
- Select dimension value combinations via checkboxes (e.g., check “XXL” row and all color columns) to generate all child SKUs in one operation.
- The system auto-generates SKUs using the configured SKU pattern, assigns the parent’s base price and cost as defaults, and sets initial inventory to zero.
- Staff can review generated variants before confirming creation.
Matrix Import via CSV:
- Upload a CSV file with columns matching dimension values and data fields.
- CSV format:
dimension_1_value, dimension_2_value, price, cost, barcode - The system validates all rows, reports errors (duplicate barcodes, invalid dimension values, non-numeric prices), and applies valid rows in batch.
- Supports both CREATE (new variants) and UPDATE (existing variants matched by dimension values) modes.
Business Rules:
- Matrix view is available only for products with exactly 2 variant dimensions. Products with 1 or 3 dimensions use the standard list editor.
- Cells for non-existent variant combinations display as empty with a “+” icon to create that specific variant on demand.
- Deleting a variant from the matrix is only permitted when the variant has zero inventory across all locations and no open orders referencing it.
3.2 Product Lifecycle
Scope: Managing products from initial creation through active selling, discontinuation, archival, and potential re-creation from archived templates. The lifecycle enforces data completeness before publishing and protects against premature archival while stock remains on hand.
3.2.1 Product Lifecycle State Machine
stateDiagram-v2
[*] --> DRAFT: Product Created
DRAFT --> ACTIVE: Publish
ACTIVE --> DISCONTINUED: Discontinue
DISCONTINUED --> ACTIVE: Reactivate
DISCONTINUED --> ARCHIVED: Archive (stock = 0 all locations)
ARCHIVED --> DRAFT: Clone as New (new SKU)
note right of DRAFT
Incomplete product
Not visible at POS
No sales allowed
end note
note right of ACTIVE
Fully configured
Available for sale
Visible at POS
end note
note right of DISCONTINUED
Sell-through mode
No restock / no new POs
Still sellable until stock = 0
end note
note right of ARCHIVED
Read-only historical record
Not visible at POS
Can clone to create new product
end note
3.2.2 Lifecycle Transition Rules
| Transition | Preconditions | Actions | Post-Conditions |
|---|---|---|---|
| Draft -> Active (Publish) | Name is set, SKU is set, at least one price assigned, category assigned | Set published_at timestamp, make visible at POS, enable inventory tracking | Product appears in POS search and barcode lookup |
| Active -> Discontinued (Discontinue) | Product must be in ACTIVE status | Set discontinued_at timestamp, flag as sell-through only, block from new purchase orders, remove from “New Arrivals” collections | Product remains sellable but will not be restocked |
| Discontinued -> Active (Reactivate) | Product must be in DISCONTINUED status | Clear discontinued_at, re-enable for purchase orders | Product fully available for sale and restocking |
| Discontinued -> Archived (Archive) | qty_on_hand = 0 at ALL locations (stores + warehouse) | Set archived_at timestamp, remove from POS visibility, mark record read-only | Product preserved for historical reporting but invisible to POS operations |
| Archived -> Draft (Clone as New) | Source product must be in ARCHIVED status | Create new product record with new auto-generated SKU, copy all fields except identity and timestamps, set status to DRAFT | New draft product exists independently; original archive unchanged |
Business Rules:
- A product cannot be archived while any location holds stock. The system checks
qty_on_handacross all locations before allowing the transition. - Cloning an archived product does NOT restore the original. It creates a new, separate product that can be edited independently.
- Discontinued products continue to appear in POS barcode lookup and search so that remaining stock can be sold.
3.3 Pricing Engine
Scope: Managing product pricing across channels, customer groups, and promotional periods. The engine supports centralized pricing with cascading overrides, named price books, four promotion types, formal markdown workflows, and best-price conflict resolution. Every price determination is auditable – the system logs which rules were evaluated, which rule won, and the final price applied.
3.3.1 Price Hierarchy
Centralized pricing with cascading override resolution. When determining the price for a product at checkout, the system evaluates five priority levels and applies the highest-priority match:
| Priority | Level | Source | Description |
|---|---|---|---|
| 1 (Highest) | Manual Override | Staff at POS | Staff manually enters a price during the sale. Requires reason selection and manager PIN if discount exceeds configurable threshold |
| 2 | Active Promotion | Promotion Engine | Scheduled or triggered promotions currently in effect for this product |
| 3 | Price Book | Price Book Engine | Customer-group-specific or date-bound pricing from a named price list |
| 4 | Channel Price | Channel Configuration | Per-channel pricing: In-Store, Online, or Wholesale |
| 5 (Lowest) | Global Default | Product Record | The base_price field from the product data model |
flowchart TD
A[Start: Determine Price] --> B{Manual Override?}
B -->|Yes| C[Use Manual Override Price]
B -->|No| D{Active Promotion?}
D -->|Yes| E[Use Promotion Price]
D -->|No| F{Price Book Match?}
F -->|Yes| G[Use Price Book Price]
F -->|No| H{Channel Price Set?}
H -->|Yes| I[Use Channel Price]
H -->|No| J[Use base_price from Product]
C --> K[Log Price Resolution to Audit Trail]
E --> K
G --> K
I --> K
J --> K
K --> L[Return Final Price]
3.3.2 Price Books
Scope: Named price lists that override global pricing for specific customer groups, channels, or date ranges. Price books enable scenarios such as wholesale pricing for B2B customers, employee discount pricing, and seasonal price adjustments without modifying the base product price.
Price Book Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Display name (e.g., “Wholesale”, “Employee Discount”, “Holiday Special”) |
description | String(500) | No | Purpose and scope of this price book |
customer_group_id | UUID | No | FK to customer_groups – restrict this book to a specific customer group. NULL = any customer |
channel | Enum | No | Restrict to channel: IN_STORE, ONLINE, WHOLESALE. NULL = all channels |
start_date | DateTime | Yes | When this price book becomes active |
end_date | DateTime | Yes | When this price book expires |
is_active | Boolean | Yes | Master toggle (default: true). Allows manual deactivation before end_date |
priority | Integer | Yes | When multiple price books match, the highest priority value wins (default: 10) |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
created_by | UUID | Yes | User who created the price book |
Price Book Entry Table
| Field | Type | Required | Description |
|---|---|---|---|
price_book_id | UUID | Yes | FK to price_books table |
product_id | UUID | Yes | FK to products table (parent or standard product) |
override_price | Decimal(10,2) | Yes | The price to charge when this price book is active |
override_cost | Decimal(10,2) | No | Optional cost override for margin reporting accuracy |
Business Rules:
- Exclusivity: Only one price book can be active per customer group per channel at any given time. If a new price book overlaps with an existing active book for the same group/channel combination, the system rejects creation and prompts the user to deactivate the conflicting book first.
- No stacking: Price books do not stack. When multiple price books match (e.g., one by customer group and one by channel), only the book with the highest
priorityvalue applies. - Audit: Every price book activation, deactivation, and entry modification is logged with the acting user and timestamp.
- Bulk entry: Staff can import price book entries via CSV with columns:
sku, override_price, override_cost. The system validates SKU existence and numeric formats before applying.
3.3.3 Promotions
Scope: Four promotion types that cover the most common retail discount scenarios. Promotions can be automatic (applied when conditions are met) or code-based (requiring manual entry at POS).
Promotion Types
| Type | Description | Key Fields |
|---|---|---|
| Basic Discount | Percentage or fixed amount off a single item | discount_type (PERCENT / AMOUNT), discount_value, min_qty (optional), max_uses (optional) |
| Tiered / Volume | Quantity breaks – price per unit decreases as quantity increases | tiers[] array of {min_qty, max_qty, price_per_unit} |
| BOGO / Cross-Item | Buy X get Y at a discount | buy_product_id, buy_qty, get_product_id, get_qty, get_discount_type, get_discount_value |
| Scheduled / Automatic | Activate on date/time without manual intervention | schedule_type (DATE_RANGE / RECURRING), start_datetime, end_datetime, recurrence_pattern |
Tiered / Volume Example:
| Quantity | Price Per Unit |
|---|---|
| 1 | $10.00 |
| 3+ | $8.00 |
| 10+ | $6.00 |
BOGO / Cross-Item Example:
- Buy 2 Shirts (
buy_product_id= shirts category,buy_qty= 2), Get 1 Belt 50% off (get_product_id= belts category,get_qty= 1,get_discount_type= PERCENT,get_discount_value= 50)
Scheduled / Automatic Example:
- Happy Hour:
schedule_type= RECURRING,recurrence_pattern= “MON-FRI 15:00-17:00” – 15% off all drinks every weekday from 3 PM to 5 PM
Promotion Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Display name (e.g., “Summer BOGO”, “Volume Discount”) |
type | Enum | Yes | BASIC, TIERED, BOGO, SCHEDULED |
stackable | Boolean | Yes | Whether this promotion can combine with other promotions (default: false) |
exclusive | Boolean | Yes | When true, this promotion replaces ALL other pricing – no other rules evaluated (default: false) |
product_scope | Enum | Yes | ALL, CATEGORY, PRODUCT_LIST |
product_ids[] | UUID[] | No | Applicable product IDs when scope = PRODUCT_LIST |
category_ids[] | UUID[] | No | Applicable category IDs when scope = CATEGORY |
start_date | DateTime | Yes | Earliest date promotion can activate |
end_date | DateTime | Yes | Latest date promotion remains active |
is_active | Boolean | Yes | Master toggle (default: false – starts as draft) |
priority | Integer | Yes | Tiebreaker when multiple promotions match (higher wins, default: 10) |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
created_by | UUID | Yes | User who created the promotion |
Promotion Lifecycle
stateDiagram-v2
[*] --> DRAFT : Create Promotion
DRAFT --> SCHEDULED : Set dates & activate
DRAFT --> CANCELLED : Cancel before scheduling
SCHEDULED --> ACTIVE : Start date reached
SCHEDULED --> CANCELLED : Cancel before start
ACTIVE --> EXPIRED : End date passed
ACTIVE --> CANCELLED : Manually deactivate
CANCELLED --> [*]
EXPIRED --> [*]
Lifecycle Transition Rules:
- DRAFT: Promotion is being configured. Not visible to POS. Can be edited freely.
- SCHEDULED: Promotion is locked for editing (except cancellation). Awaiting start date.
- ACTIVE: Promotion is live. Applied automatically or via code at POS. Only the
is_activetoggle andend_datecan be modified. - EXPIRED: Terminal state. Promotion has passed its end date. Preserved for reporting. Cannot be reactivated – must be cloned to create a new promotion.
- CANCELLED: Terminal state. Promotion was manually stopped. Preserved for reporting.
3.3.4 Markdown & Clearance Management
Scope: Formal workflows for reducing prices, tracking clearance merchandise, and handling end-of-life write-offs. Every price reduction is accountable – the system enforces approval workflows and logs all changes for audit.
Markdown Request Workflow
sequenceDiagram
autonumber
participant S as Staff
participant UI as Admin Portal
participant API as Backend
participant M as Manager
participant DB as DB
participant POS as POS Client
Note over S, POS: Markdown Request Workflow
S->>UI: Create Markdown Request
Note right of UI: Product, new_price, effective_date, reason
UI->>API: POST /markdowns
API->>DB: Save request (status: PENDING)
API-->>UI: Request #MD-001 created
API->>M: Notification: Markdown request pending
M->>UI: Review request details
Note right of M: Current price, proposed price,<br/>margin impact, reason
alt Approved
M->>API: PATCH /markdowns/MD-001 {status: APPROVED}
API->>DB: Update status, schedule price change
Note right of DB: effective_date triggers price update
DB-->>API: Price updated on effective_date
API->>POS: Push updated price to all terminals
API->>DB: Log to price_audit_trail
Note right of DB: who, when, old_price,<br/>new_price, reason, approval_id
else Rejected
M->>API: PATCH /markdowns/MD-001 {status: REJECTED, reject_reason: "..."}
API-->>S: Notification: Markdown rejected
end
Manual Price Changes:
- Staff with appropriate permissions can change a product’s price directly without the formal markdown workflow.
- Every manual change requires selecting a reason from a configurable list:
Damaged,Price Match,Manager Discretion,Competitive Adjustment,Cost Change,Seasonal Reduction,Error Correction. - All manual changes are logged to the price audit trail:
user_id,timestamp,product_id,old_price,new_price,reason,location_id.
Automatic Markdown Rules:
- Configurable rules engine that evaluates nightly and flags or automatically applies markdowns:
| Rule | Condition | Action |
|---|---|---|
| Slow Mover | No sale in X days AND stock > Y units | Reduce price by Z% |
| Aging Inventory | days_since_receive > N AND stock > threshold | Move to clearance collection, apply clearance_price |
| Season End | Season status = CLEARANCE AND product.season_id matches | Apply configured season clearance discount |
- Automatic rules can be configured to either flag for review (creates a pending markdown request) or apply immediately (executes the price change and logs it as auto-rule).
Liquidation / Write-Off:
- When a product cannot sell at any price (damaged beyond sale, expired, recalled), staff create a write-off record.
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | Product being written off |
quantity | Integer | Yes | Number of units written off |
write_off_value | Decimal(10,2) | Yes | Total cost value of written-off inventory |
reason | Enum | Yes | DAMAGED, EXPIRED, RECALLED, SHRINKAGE, OBSOLETE |
approved_by | UUID | Yes | Manager who authorized the write-off |
location_id | UUID | Yes | Location where inventory is removed |
notes | String | No | Additional context |
created_at | DateTime | Yes | Timestamp of write-off |
Clearance Rack Tracking:
- Each product can be flagged as clearance on a per-location basis.
- Clearance fields:
is_clearance(Boolean),clearance_price(Decimal),clearance_started_at(DateTime). - Products flagged as clearance appear in a dedicated “Clearance” collection on the POS browse screen and can be filtered in reports.
- Clearance pricing takes precedence over the product’s
base_pricebut is overridden by active promotions and manual overrides per the standard price hierarchy.
3.3.5 Conflict Resolution
Scope: Determining the final price when multiple pricing rules match the same product at checkout. The resolution algorithm ensures predictability, transparency, and customer-friendly outcomes.
Resolution Algorithm:
-
Check exclusivity: If any matched promotion has
exclusive = true, use ONLY that promotion. If multiple exclusive promotions match, the one with the highestprioritywins. All other pricing rules are ignored. -
If no exclusive promotion: Apply the best price for the customer (lowest final price) from among:
- The winning price book entry (if any)
- The winning non-exclusive promotion (if any)
- The channel price (if set)
- The base price
-
Stackable promotions: Stackable promotions combine ONLY when both are marked
stackable = true. The combined discount must not exceed the tenant’smax_discount_percentsetting (default: 75%). If stacking would exceed the cap, the system applies the single best promotion instead. -
Manual override: A manual override entered by staff at the POS always wins regardless of all other rules. It bypasses the algorithm entirely.
Conflict Resolution Flowchart:
flowchart TD
A[Checkout: Evaluate Price] --> B[Gather all matching rules]
B --> C{Any exclusive promotion?}
C -->|Yes| D[Use highest-priority exclusive promo]
C -->|No| E{Multiple stackable promos?}
E -->|Yes| F{Combined discount <= max_discount_percent?}
F -->|Yes| G[Apply stacked promotions]
F -->|No| H[Apply single best promotion]
E -->|No| I[Compare: Promo vs Price Book vs Channel vs Base]
I --> J[Use lowest final price for customer]
D --> K[Log: rules evaluated, winner, reason]
G --> K
H --> K
J --> K
Audit: Every price resolution is logged with the following data:
sale_id,line_item_id,product_idrules_evaluated[]– list of all pricing rules that were consideredwinning_rule_idandwinning_rule_type(MANUAL / PROMOTION / PRICE_BOOK / CHANNEL / BASE)original_price(base_price) andfinal_pricetotal_discount_amountandtotal_discount_percenttimestamp
3.3.6 Reports: Pricing
| Report | Purpose | Key Data Fields |
|---|---|---|
| Price Book Usage | Track which price books are active and how often applied | Price book name, times applied, avg discount %, revenue impact, active date range |
| Promotion Performance | Measure promotion effectiveness against business goals | Promo name, type, redemptions, revenue lift vs. pre-promo period, margin impact, cost of discount given |
| Markdown History | Track all price reductions with full accountability | Product SKU, old price, new price, reason, who requested, who approved, when, approval status |
| Margin Impact Analysis | Understand how pricing decisions affect profitability | Product/category, base margin %, discounted margin %, volume at each price point, total margin dollars lost/gained |
| Conflict Resolution Log | Audit which pricing rules win at checkout | Transaction ID, product, rules evaluated, winning rule, final price, discount applied |
| Clearance Tracking | Monitor clearance inventory and sell-through | Product, original price, clearance price, days on clearance, units remaining, sell-through % |
3.4 Barcode Management
Scope: Supporting multiple barcode formats per product, auto-generating internal barcodes when no manufacturer code exists, enforcing uniqueness per tenant, and enabling fast barcode-to-product lookup from any scanner or manual entry.
3.4.1 Barcode Types
| Type | Format | Source | Example | Use Case |
|---|---|---|---|---|
| UPC-A | 12-digit numeric | Manufacturer-assigned | 012345678901 | Standard North American retail products |
| EAN-13 | 13-digit numeric | Manufacturer-assigned (EU/International) | 4006381333931 | International products, European imports |
| Internal | Configurable prefix + auto-increment sequence | System-generated | INT-000001, INT-000002 | Products without manufacturer barcodes (custom items, local goods) |
| Alternate | Any format, free-form | Manual entry | VENDOR-SKU-X99, OLD-UPC-123 | Vendor SKUs, legacy barcodes, secondary identifiers |
3.4.2 Barcode Features
- Multiple barcodes per product: Each product has one primary barcode and unlimited alternate barcodes. Scanning ANY barcode (primary or alternate) resolves to the same product.
- Auto-generate internal barcode: When a product is created without a manufacturer barcode, the system auto-generates an internal barcode using the tenant’s configured prefix and the next available sequence number.
- Uniqueness enforced per tenant: No two products within the same tenant can share a barcode (primary or alternate). The system rejects duplicates at creation and import time.
- Universal scan resolution: POS barcode lookup checks the primary barcode first, then searches alternate barcodes. The lookup returns the same product regardless of which barcode was scanned.
- Bulk barcode import via CSV: Staff can upload a CSV file mapping barcodes to existing SKUs. The system validates uniqueness, reports conflicts, and applies valid mappings in batch.
3.4.3 Barcode Lookup Flow
sequenceDiagram
autonumber
participant U as Staff
participant SC as Scanner
participant UI as POS UI
participant API as Backend
participant DB as DB
Note over U, DB: Barcode Scan to Product Lookup
alt Barcode Scanner
U->>SC: Scan Item Barcode
SC->>UI: Barcode Value (e.g., "012345678901")
else Manual Entry
U->>UI: Type Barcode / SKU
end
UI->>API: POST /products/barcode-lookup
Note right of API: {barcode: "012345678901", tenant_id: "..."}
API->>DB: SELECT * FROM products WHERE barcode = ?
Note right of DB: Check primary barcode first
alt Primary Barcode Match
DB-->>API: Product Found
else No Primary Match
API->>DB: SELECT * FROM alternate_barcodes WHERE barcode = ?
Note right of DB: Search alternate barcodes
alt Alternate Barcode Match
DB-->>API: Product Found (via alternate)
else No Match Found
DB-->>API: No Results
API-->>UI: "Product Not Found"
UI-->>U: "No product matches this barcode"
Note right of UI: Option: Create new product or re-scan
end
end
API-->>UI: Return Product Data
UI-->>U: Display: Name, SKU, Price, Stock Level
Note right of UI: Product ready to add to cart
3.4.4 Reports: Barcode Management
| Report | Purpose | Key Data Fields |
|---|---|---|
| Barcode Coverage Report | Identify products missing barcodes | Products without primary barcode, products with only internal barcodes, total coverage % |
| Barcode Scan Failure Log | Track unrecognized scans | Scanned value, timestamp, terminal, resolution (created new / manual lookup / abandoned) |
| Duplicate Barcode Audit | Detect and prevent barcode conflicts | Barcode value, conflicting products, resolution status |
3.5 Categories, Seasons & Collections
Scope: Organizing products into a navigable hierarchy for POS browsing, reporting, and rule application (tax defaults, commission rates). The system supports up to four levels of nesting, freeform tags, named collections, auto-tagging rules, formal buying seasons with lifecycle management, and multi-dimensional reporting hierarchies for financial analysis.
3.5.1 Category Hierarchy
The catalog supports a 4-level hierarchy. Products are assigned to the most specific level applicable.
Level 1: Department
|-- Level 2: Category
|-- Level 3: Subcategory
|-- Level 4: Sub-subcategory (max depth)
Example:
Men's Apparel (Department)
|-- Tops (Category)
| |-- T-Shirts (Subcategory)
| | |-- Graphic Tees (Sub-subcategory)
| | |-- Plain Tees (Sub-subcategory)
| |-- Dress Shirts (Subcategory)
| |-- Polos (Subcategory)
|-- Bottoms (Category)
| |-- Jeans (Subcategory)
| |-- Chinos (Subcategory)
|-- Outerwear (Category)
|-- Jackets (Subcategory)
|-- Coats (Subcategory)
Accessories (Department)
|-- Bags (Category)
|-- Hats (Category)
|-- Jewelry (Category)
Services (Department)
|-- Alterations (Category)
|-- Gift Services (Category)
3.5.2 Tags & Collections
Tags: Unlimited freeform tags can be applied to any product. Tags are tenant-scoped and support search filtering.
| Feature | Description | Example |
|---|---|---|
| Freeform Tags | Any text, lowercase normalized, no hierarchy | summer, bestseller, organic, limited-edition |
| Named Collections | Curated product groups, manual or rule-based | “Summer Essentials”, “Staff Picks”, “New Arrivals” |
| Manual Collections | Staff manually adds/removes products | “Staff Picks” – staff curates the list |
| Rule-Based Collections | Auto-populated based on conditions | “New Arrivals” = products where published_at is within last 30 days |
| Auto-Tagging Rules | IF condition THEN add tag automatically | IF category = “Outerwear” AND created_at within last 14 days THEN add tag new-outerwear |
3.5.3 Category Features
- Drag-and-drop reordering: Staff can reorder categories and products within categories via drag-and-drop in the catalog management UI.
- Category-level default tax code: Each category can specify a default tax code (e.g., “clothing_exempt”, “grocery_food”). Products inherit the category tax code unless overridden at the product level.
- Category-level default commission rate: Each category can specify a default commission percentage. Used for commission calculation unless overridden at the product level.
- Bulk move products: Staff can select multiple products and move them to a different category in a single action.
- Category image/icon: Each category can have an assigned image or icon for display in POS browse mode and kiosk interfaces.
3.5.4 Reports: Category Management
| Report | Purpose | Key Data Fields |
|---|---|---|
| Category Sales Report | Revenue breakdown by category | Category path, product count, units sold, revenue, avg margin |
| Uncategorized Products | Find products missing a category | Product SKU, name, created date, status |
| Category Depth Report | Audit hierarchy usage | Depth level, category count, product count per level |
3.5.5 Formal Seasons
Scope: Named buying seasons with lifecycle dates that track merchandise from initial buy planning through active selling, clearance, and close-out. Seasons provide a temporal dimension to inventory analysis – enabling sell-through tracking, carryover identification, and season-over-season comparison.
Season Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Display name (e.g., “Spring 2026”, “Holiday 2025”, “Back to School 2026”) |
start_date | Date | Yes | Season merchandise begins arriving / becomes active |
end_date | Date | Yes | Season officially closes – remaining inventory is carryover |
status | Enum | Yes | PLANNING, ACTIVE, CLEARANCE, CLOSED |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Example Seasons:
| Season Name | Start Date | End Date | Status |
|---|---|---|---|
| Spring 2026 | 2026-02-01 | 2026-07-31 | PLANNING |
| Holiday 2025 | 2025-10-15 | 2026-01-15 | CLEARANCE |
| Fall 2025 | 2025-08-01 | 2025-12-31 | CLOSED |
| Summer 2026 | 2026-05-01 | 2026-09-30 | PLANNING |
Season Lifecycle State Machine
stateDiagram-v2
[*] --> PLANNING : Create Season
PLANNING --> ACTIVE : Season start date reached
ACTIVE --> CLEARANCE : Triggered by staff or auto-rule
CLEARANCE --> CLOSED : Season end date passed
PLANNING --> CLOSED : Cancel season (no products received)
CLOSED --> [*]
Lifecycle Transition Rules:
- PLANNING: Season is being prepared. Products can be assigned to this season. Purchase orders can reference this season. No products are expected on the sales floor yet.
- ACTIVE: Season merchandise is on the sales floor and selling. Transition occurs automatically when
start_dateis reached, or manually by staff. Products assigned to this season appear in season-filtered reports. - CLEARANCE: Triggered manually by a buyer/manager OR by an automatic rule (e.g., “30 days before end_date, move to CLEARANCE”). Products in this season become eligible for automatic markdown rules. The clearance collection auto-populates with this season’s products.
- CLOSED: Terminal state. Season has ended. Remaining inventory is flagged as carryover. No further price changes tied to this season. Season data is preserved for historical reporting.
Product-Season Assignment:
- Products are assigned to seasons via the
season_idFK on the product record. - A product can belong to at most ONE season. Products with
season_id = NULLare year-round core items not tied to any season. - When a season transitions to CLOSED, products still assigned to it can be reassigned to a new season (carryover) or have their
season_idset to NULL (promoted to core).
3.5.6 Reporting Dimensions
Scope: Structured classification hierarchies used exclusively for financial reporting and buying analysis. Reporting dimensions are separate from display categories – a product’s display category determines where it appears in the POS browse screen, while reporting dimensions determine how it appears in financial reports, open-to-buy analysis, and margin summaries.
Brand
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Brand name (e.g., “Nike”, “Levi’s”, “Nexus Premier”) |
logo_url | String(500) | No | URL to brand logo image |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
- Products are assigned to brands via the
brand_idFK on the product record. - Brand is distinct from Vendor. Nike the brand (what the customer sees on the label) is different from Nike Direct the vendor (the company you send purchase orders to). A single brand may be sourced from multiple vendors. A single vendor may supply multiple brands.
Merchandise Hierarchy
A 3-level reporting hierarchy independent of display categories:
Department (Level 1)
|-- Class (Level 2)
|-- Subclass (Level 3)
Example:
Footwear (Department)
|-- Athletic Shoes (Class)
| |-- Running (Subclass)
| |-- Basketball (Subclass)
| |-- Training (Subclass)
|-- Casual Shoes (Class)
| |-- Sneakers (Subclass)
| |-- Loafers (Subclass)
|-- Dress Shoes (Class)
|-- Oxfords (Subclass)
|-- Derby (Subclass)
Merchandise Hierarchy Fields on Product:
| Field | Type | Required | Description |
|---|---|---|---|
merch_department_id | UUID | No | FK to merchandise_departments table |
merch_class_id | UUID | No | FK to merchandise_classes table (must be child of selected department) |
merch_subclass_id | UUID | No | FK to merchandise_subclasses table (must be child of selected class) |
- The merchandise hierarchy enables financial reporting by buying category rather than display category. A “Galaxy V-Neck Tee” might display under “Men’s > Casual > T-Shirts” but report under “Apparel > Knit Tops > V-Necks” for buying analysis.
- Merchandise hierarchy assignment is optional. Products without a merchandise hierarchy assignment are grouped under “Unclassified” in financial reports.
Custom Dimensions
Tenant-defined reporting dimensions for analysis needs beyond brand and merchandise hierarchy.
| Field | Type | Required | Description |
|---|---|---|---|
dimension_id | UUID | Yes | Primary key |
name | String(100) | Yes | Dimension name (e.g., “Buyer”, “Margin Tier”, “Velocity Class”) |
values[] | String[] | Yes | List of allowed values (e.g., [“Sarah”, “Mike”, “Unassigned”] for Buyer dimension) |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
Product-Dimension Junction Table:
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | FK to products table |
dimension_id | UUID | Yes | FK to custom_dimensions table |
dimension_value | String | Yes | Selected value (must be one of values[] from the dimension definition) |
Example Custom Dimensions:
| Dimension | Allowed Values | Purpose |
|---|---|---|
| Buyer | Sarah, Mike, Jenny, Unassigned | Track which buyer selected this product for reporting by buyer performance |
| Margin Tier | High (>60%), Medium (40-60%), Low (<40%) | Quick filtering for margin-focused analysis |
| Velocity Class | A (top 20%), B (middle 60%), C (bottom 20%) | ABC analysis classification for inventory planning |
| Sourcing Region | Domestic, Asia, Europe, South America | Supply chain analysis and lead time planning |
3.5.7 Reports: Seasons & Dimensions
| Report | Purpose | Key Data Fields |
|---|---|---|
| Season Sell-Through | Track sell-through rate per season to evaluate buying accuracy | Season name, total units received, total units sold, sell-through %, revenue, avg margin % |
| Season Carryover | Identify unsold season inventory for markdown or transfer decisions | Season name, product SKU, product name, qty remaining, original cost, current retail value, days since season close |
| Brand Performance | Revenue and margin analysis by brand for vendor negotiation and buying decisions | Brand name, product count, units sold, revenue, margin %, return rate, avg selling price |
| Merchandise Hierarchy Report | Financial reporting by Dept/Class/Subclass for open-to-buy and assortment planning | Department, Class, Subclass, revenue, margin %, inventory value at cost, inventory turns, weeks of supply |
| Custom Dimension Report | Flexible analysis by any tenant-defined dimension | Dimension name, dimension value, product count, revenue, margin %, units sold, avg price |
3.6 Multi-Channel Management
Scope: Managing product visibility, inventory allocation, and pricing across multiple sales channels (In-Store POS, Online/Shopify, Wholesale). Each product can be configured independently per channel, enabling retailers to control where products appear, how much stock each channel can sell, and at what price.
Cross-Reference: See Module 4, Section 4.2 for inventory allocation across channels.
3.6.1 Channel Definition
The system ships with three built-in channels. Tenants can define additional custom channels to match their sales operations.
| Channel | Type | Default | Description |
|---|---|---|---|
| IN_STORE | PHYSICAL | Yes | Point-of-sale transactions at physical store locations |
| ONLINE | DIGITAL | No | E-commerce sales via Shopify or other web storefront |
| WHOLESALE | B2B | No | Bulk/wholesale orders from business customers |
Channel Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
name | String | Yes | Display name (e.g., “In-Store POS”, “Shopify Online”, “Wholesale Portal”) |
type | Enum | Yes | PHYSICAL, DIGITAL, B2B |
is_default | Boolean | Yes | Whether this is the default channel for new products (one per tenant) |
is_system | Boolean | Yes | System-defined channels cannot be deleted |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
3.6.2 Channel Visibility Controls
Each product is independently toggled for visibility on each channel. Visibility can be immediate or scheduled with date windows.
Channel Visibility Data Model
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | Reference to product |
channel_id | UUID | Yes | Reference to channel |
is_visible | Boolean | Yes | Whether the product appears on this channel |
available_from | DateTime | No | Scheduled visibility start (null = immediately visible when toggled on) |
available_until | DateTime | No | Scheduled visibility end (null = indefinite) |
tenant_id | UUID | Yes | Owning tenant |
Business Rules:
- A product must be visible on at least one channel to remain in
ACTIVElifecycle status. If a product is removed from all channels, the system warns: “Product will be hidden from all sales channels. Consider setting to DISCONTINUED.” - Scheduled visibility windows are evaluated in real time. A product with
available_fromin the future does not appear on the channel until that timestamp. - Bulk operations: staff can select multiple products and toggle channel visibility in batch via the catalog management UI.
Channel Visibility Sequence
sequenceDiagram
autonumber
participant U as Staff
participant UI as Catalog UI
participant API as Backend
participant DB as DB
Note over U, DB: Single Product Channel Toggle
U->>UI: Open Product Detail
UI->>API: GET /products/{id}/channels
API-->>UI: Return Channel Visibility Settings
U->>UI: Toggle "Online" Channel ON
U->>UI: Set available_from "2026-03-01"
U->>UI: Click "Save"
UI->>API: PATCH /products/{id}/channels
API->>DB: Upsert Channel Visibility Record
API-->>UI: Channel Updated
Note over U, DB: Bulk Channel Toggle
U->>UI: Select 25 Products (Checkbox)
U->>UI: Click "Bulk Actions" -> "Channel Visibility"
UI-->>U: Show Channel Toggle Panel
U->>UI: Toggle "Wholesale" ON for All Selected
U->>UI: Click "Apply"
UI->>API: POST /products/bulk-channel-update
Note right of API: {product_ids: [...], channel_id: "...", is_visible: true}
API->>DB: Upsert 25 Channel Visibility Records
API-->>UI: "25 products updated"
3.6.3 Channel Inventory Allocation
Each tenant selects one of two inventory allocation modes. The mode applies globally per tenant.
| Mode | Description | Use Case |
|---|---|---|
| Shared Pool (Default) | All channels sell from the same inventory pool. An online sale decrements the same stock as an in-store sale. | Small-to-medium retailers with unified stock |
| Dedicated Allocation | Reserve specific quantities per channel per location. Each channel has its own available pool. | Large retailers needing channel-specific stock control |
Dedicated Allocation Data Model
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | Reference to product |
location_id | UUID | Yes | Reference to store/warehouse location |
channel_id | UUID | Yes | Reference to channel |
allocated_qty | Integer | Yes | Quantity reserved for this channel at this location |
sold_qty | Integer | Yes | Quantity sold through this channel (starts at 0) |
available_qty | Integer | Computed | Calculated: allocated_qty - sold_qty |
tenant_id | UUID | Yes | Owning tenant |
last_updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Sum of
allocated_qtyacross all channels at a location cannot exceed total physical inventory at that location. - When one channel sells out (available_qty = 0), the system flags the product as “Channel Stockout” in the dashboard. Staff can manually release allocation from another channel to replenish.
- Auto-reallocation is not automatic by default. Staff must approve reallocation to prevent unintended stock shifts.
- In Shared Pool mode, the dedicated allocation table is not used. All channels reference the location’s total
qty_on_hand.
3.6.4 Channel Pricing
Each product can carry a different price per channel. Channel pricing ties into the Price Tier hierarchy defined in the pricing engine.
Channel Pricing Data Model
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | Reference to product |
channel_id | UUID | Yes | Reference to channel |
channel_price | Decimal(10,2) | No | Override price for this channel (null = use base_price) |
channel_compare_at_price | Decimal(10,2) | No | “Was” price for this channel (used for strikethrough display) |
tenant_id | UUID | Yes | Owning tenant |
Business Rules:
- If no
channel_priceis set for a product-channel combination, the system falls through to the product’s globalbase_price. - Channel pricing is evaluated BEFORE customer price tiers. The resolution order is: Channel Price -> Base Price -> Price Tier Override.
- Wholesale channel pricing typically represents a cost-plus markup and may be lower than in-store pricing.
3.6.5 Reports: Multi-Channel
| Report | Purpose | Key Data Fields |
|---|---|---|
| Channel Sales Comparison | Compare revenue performance across channels | Channel, units sold, revenue, gross margin, avg order value, period |
| Channel Inventory Status | Stock availability by channel | Product, channel, allocated qty, available qty, sold qty, stockout risk flag |
| Channel Visibility Audit | Identify products missing from channels | Product, channels currently visible, channels not visible, last change date, recommended action |
| Channel Price Variance | Price differences across channels for same product | Product, in-store price, online price, wholesale price, variance % |
3.7 Shopify Integration
MOVED TO MODULE 6: This section has been consolidated into Module 6: Integrations & External Systems, Section 6.3 (Shopify Integration). The full Shopify integration specification – including sync modes, field-level ownership, conflict resolution, sync constraints, GraphQL API preference, @idempotent directive, Bulk Operations API, third-party POS integration rules, omnichannel/BOPIS requirements, and hardware compatibility – is now maintained in Section 6.3.
See: Module 6, Section 6.3 for the complete Shopify integration specification.
3.8 Vendor Management
Scope: Tracking supplier information, payment terms, and the many-to-many relationship between vendors and products. A single product can be sourced from multiple vendors, each with vendor-specific cost, SKU, and lead time.
Cross-Reference: See Module 4, Section 4.3 for purchase orders and Section 4.9 for vendor RMA.
Cross-Reference: See Module 5, Section 5.19 for supplier payment terms and lead time configuration.
3.8.1 Vendor Data Model
| Group | Field | Type | Required | Description |
|---|---|---|---|---|
| Identity | id | UUID | Yes | Primary key, system-generated |
name | String | Yes | Vendor company name | |
code | String | Yes | Short unique code (e.g., “NIKE”, “LEVI”) | |
tax_id | String | No | Vendor tax identification number | |
| Contact | email | String | No | Primary contact email |
phone | String | No | Primary phone number | |
address | Object | No | Full mailing address (street, city, state, zip, country) | |
contact_person | String | No | Name of primary contact | |
| Terms | payment_terms | Enum | Yes | NET_30, NET_60, NET_90, COD, PREPAID |
currency | String | Yes | Default currency code (e.g., “USD”) | |
minimum_order | Decimal(10,2) | No | Minimum order amount required | |
| Logistics | default_lead_time_days | Integer | No | Standard delivery lead time in days |
preferred_carrier | String | No | Default shipping carrier | |
| Status | status | Enum | Yes | ACTIVE, INACTIVE |
| Timestamps | created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
3.8.2 Vendor-Product Relationship
Vendors and products share a many-to-many relationship through the vendor_product junction table. Each link carries vendor-specific pricing, SKU mapping, and lead time overrides.
| Field | Type | Required | Description |
|---|---|---|---|
vendor_id | UUID | Yes | Reference to vendor |
product_id | UUID | Yes | Reference to product |
vendor_sku | String | No | The vendor’s own SKU for this product |
vendor_cost | Decimal(10,2) | Yes | Purchase cost from this vendor |
vendor_barcode | String | No | Vendor-specific barcode (can differ from product barcode) |
is_primary_vendor | Boolean | Yes | Whether this is the preferred vendor for this product (one per product) |
lead_time_override_days | Integer | No | Vendor-product specific lead time (overrides vendor default) |
minimum_order_qty | Integer | No | Minimum quantity per order for this product from this vendor |
last_ordered_at | DateTime | No | Timestamp of most recent PO to this vendor for this product |
3.8.3 Vendor-Product Entity Relationship
erDiagram
VENDOR {
UUID id PK
String name
String code
String tax_id
String email
String phone
String payment_terms
String currency
Decimal minimum_order
Integer default_lead_time_days
String status
}
PRODUCT {
UUID id PK
String sku
String name
String product_type
Decimal base_price
Decimal cost
String lifecycle_status
}
VENDOR_PRODUCT {
UUID vendor_id FK
UUID product_id FK
String vendor_sku
Decimal vendor_cost
String vendor_barcode
Boolean is_primary_vendor
Integer lead_time_override_days
Integer minimum_order_qty
}
VENDOR ||--o{ VENDOR_PRODUCT : "supplies"
PRODUCT ||--o{ VENDOR_PRODUCT : "sourced from"
3.8.4 Reports: Vendor Management
| Report | Purpose | Key Data Fields |
|---|---|---|
| Vendor Product List | All products supplied by a vendor | Vendor, product SKU, vendor SKU, vendor cost, is primary, last ordered |
| Vendor Performance Report | Track delivery and quality | Vendor, PO count, on-time %, avg lead time, variance count, return rate |
| Primary Vendor Coverage | Ensure all products have a primary vendor | Products without primary vendor, products with only one vendor source |
| Vendor Cost Comparison | Compare pricing across vendors for same product | Product SKU, vendor name, vendor cost, lead time, minimum qty |
3.9 Product Search & Discovery
Scope: Enabling fast product lookup at the POS terminal through multiple search methods: full-text search, category browsing, advanced filters, favorites, and product recommendations. Search must return results in under 200ms for responsive POS operation.
3.9.1 Full-Text Search
The POS search engine indexes the following fields for fast retrieval:
| Search Field | Weight (Relevance) | Index Type | Example Match |
|---|---|---|---|
sku | 10 (highest) | Exact match | Search “BLK-TEE-001” returns exact product |
barcode / alternate_barcodes | 10 (highest) | Exact match | Search “012345678901” returns exact product |
name | 8 | Full-text, prefix | Search “Oxford” matches “Oxford Button-Down Shirt” |
brand_name | 6 | Full-text | Search “Nike” matches all Nike products |
vendor_name | 5 | Full-text | Search “Levis” matches products from Levi’s vendor |
tags[] | 4 | Exact token match | Search “bestseller” matches tagged products |
short_description | 3 | Full-text, contains | Search “moisture wicking” matches description |
long_description | 2 | Full-text, contains | Lowest priority, broad match |
custom_attributes | 3 | Key-value match | Search “organic” matches custom attribute values |
Relevance Ranking Order:
1. Exact SKU match
2. Exact barcode match (primary or alternate)
3. Name starts with search term
4. Name contains search term
5. Brand name match
6. Vendor name match
7. Tag exact match
8. Description contains search term
9. Custom attribute value match
Fuzzy Matching:
- Handles typos using Levenshtein distance (edit distance <= 2 for words >= 5 characters)
- Examples: “oxfrd” matches “Oxford”, “clasic” matches “Classic”, “niike” matches “Nike”
- Fuzzy results ranked below exact matches
- Disabled for SKU and barcode fields (exact match only)
Auto-Complete:
- Suggestions appear after 2 characters typed
- Returns top 8 suggestions combining product names, SKUs, and recent searches
- Debounced at 150ms to prevent excessive API calls
- Keyboard navigation supported (arrow keys + Enter to select)
Recent Searches:
| Field | Type | Description |
|---|---|---|
id | UUID | Primary key |
user_id | UUID | Staff member who performed the search |
search_term | String(255) | The search query text |
result_count | Integer | Number of results returned |
selected_product_id | UUID (nullable) | Product selected from results (null if no selection) |
tenant_id | UUID | Tenant scope |
created_at | DateTime | When the search was performed |
- Last 10 searches stored per user, displayed in a dropdown when the search field is focused
- Tapping a recent search re-executes the query
- Clear individual or all recent searches
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant IDX as Search Index
participant DB as DB
U->>UI: Type "oxfrd" in search bar
Note right of UI: Debounce 150ms
UI->>API: GET /products/search?q=oxfrd&limit=20
API->>IDX: Full-text query with fuzzy matching
par Search Execution
IDX->>IDX: Exact SKU/barcode check (no match)
IDX->>IDX: Full-text name search (fuzzy: "oxfrd" → "Oxford")
IDX->>IDX: Brand/vendor/tag search
IDX->>IDX: Description search
end
IDX-->>API: Ranked results (Oxford Button-Down first)
API->>DB: Fetch stock levels for results
API-->>UI: Return products with price, stock, images
UI-->>U: Display results grid (name, SKU, price, stock, image)
U->>UI: Tap "Oxford Button-Down Shirt"
UI->>UI: Add to cart
par Background
UI->>API: POST /search-history
API->>DB: Save recent search "oxfrd" for user
end
3.9.2 Category Browsing
Staff can browse the catalog visually through the 4-level category hierarchy defined in Section 3.4.
Visual Grid View:
- Categories displayed as image tiles (using
category_imagefrom Section 3.4.3) - Each tile shows: category name, product count badge, category image (or placeholder icon)
- Grid layout: configurable 3x3, 4x4, or 5x5 per screen (setting per terminal)
Drill-Down Navigation:
Department → Category → Subcategory → Products
Example:
Men's Apparel [42 items] →
Tops [28 items] →
T-Shirts [15 items] →
[Product Grid: Classic Tee, Graphic Tee, V-Neck, ...]
Breadcrumb Navigation:
Always visible at the top of the browse screen:
Home > Men's Apparel > Tops > T-Shirts
Tapping any breadcrumb segment navigates back to that level.
Sort Within Category:
| Sort Option | Direction | Default |
|---|---|---|
| Name | A-Z, Z-A | A-Z (default) |
| Price | Low-High, High-Low | – |
| Newest | Most recently published first | – |
| Best-Selling | Highest sales velocity first | – |
3.9.3 Advanced Filters
Filters are combinable using AND logic. Each active filter narrows the result set.
| Filter | Type | Values | UI Control |
|---|---|---|---|
| Price range | Range | Min-Max price (tenant currency) | Dual-handle slider with manual entry |
| Brand | Multi-select | List of all brands in tenant catalog | Searchable checkbox list |
| Vendor | Multi-select | List of all active vendors | Searchable checkbox list |
| Stock status | Multi-select | In Stock, Low Stock, Out of Stock | Checkbox group |
| Season | Multi-select | Active season tags/collections | Checkbox list |
| Size | Multi-select | Available sizes (from variant dimension values) | Checkbox grid |
| Color | Multi-select | Available colors (from variant dimension values) | Color swatch grid |
| Channel | Multi-select | In-Store, Online, Wholesale | Checkbox group |
| Lifecycle status | Multi-select | Active, Discontinued | Checkbox group |
| Category | Tree-select | Full category hierarchy | Expandable tree |
| Custom attributes | Dynamic | Based on tenant’s custom attribute definitions | Auto-generated per attribute type |
Saved Filters:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Display name (e.g., “Nike Low Stock”, “Clearance Items”) |
filter_definition | JSON | Yes | Serialized filter state: {"brand": ["Nike"], "stock_status": ["LOW_STOCK"]} |
created_by | UUID | Yes | Staff member who created the filter |
is_shared | Boolean | Yes | Whether other staff can see and use this filter (default: false) |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
- Staff can save current filter combination as a named view
- Maximum 20 saved filters per user, 50 shared filters per tenant
- Shared filters visible to all staff in the tenant
3.9.4 Quick-Add & Favorites
Favorites:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
user_id | UUID | Yes | Staff member |
product_id | UUID | Yes | Favorited product |
sort_order | Integer | Yes | Display position |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | When favorited |
- Each staff member can pin up to 50 products
- One-tap add to cart from favorites panel
- Drag-drop reordering of favorites
- Favorites persist across terminals (tied to user, not device)
Quick-Add Buttons:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
location_id | UUID | Yes | Location this configuration applies to |
product_id | UUID | Yes | Product assigned to the button |
grid_position | Integer | Yes | Position in the grid (1-20) |
button_color | String(7) | No | Hex color for button background (e.g., “#FF5733”) |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Creation timestamp |
- Configurable grid of up to 20 high-velocity items on POS home screen
- Each button displays: product image thumbnail, product name (truncated), price
- Configurable per location (e.g., store-specific quick items)
- Manager or Admin role required to configure quick-add buttons
Recent Products:
- Last 20 products viewed or added to cart, displayed in a “Recents” side panel
- Per-user, per-session (resets on logout)
- One-tap to add to cart
- Stored in local state (not persisted to database)
Department Quick-Keys:
- Configurable shortcut buttons for top-level categories (departments)
- Up to 8 department quick-keys displayed in a horizontal bar above the search field
- Tapping a quick-key navigates directly to that department’s category browse view
- Configurable per location by Manager or Admin
3.9.5 Product Substitutions & Recommendations
Out-of-Stock Alternatives:
When a scanned or searched product has zero available stock at the current location, the system presents alternatives in priority order:
| Priority | Suggestion Type | Logic | Display |
|---|---|---|---|
| 1 | Same product at another location | Query inventory where product_id matches AND qty_available > 0 at other locations | Location name, stock qty, distance (if configured) |
| 2 | Similar products | Same category_id AND base_price within +/- 20% of original AND lifecycle_status = ACTIVE AND qty_available > 0 | Product name, price, stock qty |
| 3 | Same brand alternatives | Same brand AND lifecycle_status = ACTIVE AND qty_available > 0, ordered by sales velocity | Product name, price, stock qty |
Cross-Sell / Upsell Configuration:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | Source product |
related_product_id | UUID | Yes | Related product |
relationship_type | Enum | Yes | SUBSTITUTE, CROSS_SELL, UPSELL, ACCESSORY |
priority | Integer | Yes | Display order (1 = highest priority) |
auto_generated | Boolean | Yes | true if system-generated from order history, false if manually defined |
confidence_score | Decimal(3,2) | No | For auto-generated: co-purchase frequency score (0.00-1.00) |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Creation timestamp |
Manual Relationships:
- Staff defines “related products” per product via the admin product detail page
- Relationship types: SUBSTITUTE (alternative), CROSS_SELL (complementary), UPSELL (upgrade), ACCESSORY (add-on)
- Maximum 10 manual relationships per product per type
Auto-Generated Relationships:
- Background job analyzes completed order history (rolling 90 days)
- Identifies products frequently purchased together (co-purchase frequency >= 3 occurrences)
- Generates
CROSS_SELLrelationships with confidence score based on frequency - Auto-generated relationships are refreshed weekly
- Staff can promote an auto-generated relationship to manual (persists beyond refresh)
POS Display:
- “Customers Also Bought” panel shown during checkout when cart contains products with cross-sell relationships
- Maximum 4 suggestions displayed, ordered by priority then confidence score
- Staff can dismiss suggestions or tap to add to cart
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
U->>UI: Scan barcode for "Classic Tee"
UI->>API: POST /products/barcode-lookup
API->>DB: Fetch product + inventory
DB-->>API: Product found, qty_available = 0
API-->>UI: Product data (OUT OF STOCK)
UI-->>U: "Classic Tee is out of stock at this location"
par Fetch Alternatives
API->>DB: Same product at other locations
DB-->>API: Store B: 12 units, Store C: 5 units
and
API->>DB: Similar products (same category, ±20% price)
DB-->>API: "V-Neck Tee" $27.99 (8 in stock), "Crew Neck" $31.99 (15 in stock)
and
API->>DB: Same brand alternatives
DB-->>API: "Classic Tee V2" $29.99 (20 in stock)
end
UI-->>U: Display alternatives panel
Note right of UI: 1. Same item at Store B (12) / Store C (5)
Note right of UI: 2. V-Neck Tee $27.99 (8) / Crew Neck $31.99 (15)
Note right of UI: 3. Classic Tee V2 $29.99 (20)
alt Staff selects alternative
U->>UI: Tap "V-Neck Tee"
UI->>UI: Add to cart
else Staff initiates transfer
U->>UI: Tap "Request from Store B"
UI->>API: POST /transfers/request
end
3.9.6 Reports: Search & Discovery
| Report | Purpose | Key Data Fields |
|---|---|---|
| Search Failure Log | Searches that returned zero results, identifying catalog gaps | Search term, timestamp, terminal, staff member, location |
| Top Searches | Most frequent search terms for merchandising insight | Search term, frequency, avg results returned, conversion rate (searches that led to cart add) |
| Search Conversion Funnel | Track search-to-sale effectiveness | Total searches, searches with results, searches with cart add, searches leading to sale |
| Substitute Offered vs Accepted | Effectiveness of substitution suggestions | Original product, substitute offered, offered count, accepted count, accepted %, declined % |
| Favorites Usage | Staff utilization of favorites feature | Staff member, favorite count, favorites used in sales, top 10 favorited products |
| Quick-Add Button Performance | Click-through rate of quick-add buttons | Button position, product, click count, resulting sales, revenue from quick-add |
3.10 Label & Price Tag Printing
Scope: Generating and printing barcode labels, shelf price tags, and clearance stickers from the catalog. Supports multiple label formats, batch printing, and integration with standard label printers (Zebra, DYMO, Brother).
3.10.1 Label Types
| Label Type | Content Fields | Use Case | Standard Size |
|---|---|---|---|
| Barcode Label | SKU, barcode (scannable), product name, price | Tagging individual items for checkout scanning | 50mm x 25mm |
| Shelf Price Tag | Product name, price, compare_at_price (if on sale), category, SKU | Shelf-edge display showing current price | 60mm x 40mm |
| Clearance Sticker | Markdown price, original price (strikethrough), discount %, “CLEARANCE” badge | Identifying marked-down clearance items | 40mm x 30mm |
| Variant Label | Parent name, size, color (variant dimensions), SKU, barcode (scannable), price | Per-variant tagging for variant products | 50mm x 25mm |
| Bin Label | SKU, barcode (scannable), product name, location code, bin/shelf position | Warehouse bin identification for inventory management | 100mm x 50mm |
3.10.2 Label Templates
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
name | String(100) | Yes | Template display name (e.g., “Standard Barcode 50x25”) |
label_type | Enum | Yes | BARCODE, SHELF_TAG, CLEARANCE, VARIANT, BIN |
width_mm | Integer | Yes | Label width in millimeters |
height_mm | Integer | Yes | Label height in millimeters |
layout_definition | JSON | Yes | Field positions, font sizes, barcode format, margins (see below) |
barcode_format | Enum | Yes | CODE128, EAN13, UPCA, QR_CODE |
is_default | Boolean | Yes | Default template for this label type (one per type per tenant) |
printer_language | Enum | Yes | ZPL (Zebra), DYMO_XML, BROTHER_ESC, RECEIPT_ESC_POS |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Layout Definition JSON Structure:
{
"fields": [
{ "name": "product_name", "x": 2, "y": 2, "font_size": 10, "max_width": 46, "bold": true },
{ "name": "barcode", "x": 2, "y": 10, "height": 8, "format": "CODE128" },
{ "name": "sku", "x": 2, "y": 20, "font_size": 7, "bold": false },
{ "name": "price", "x": 35, "y": 2, "font_size": 12, "bold": true, "prefix": "$" }
],
"margins": { "top": 1, "right": 1, "bottom": 1, "left": 1 }
}
Supported Printers:
| Printer | Language | Connection | Notes |
|---|---|---|---|
| Zebra ZD/ZT Series | ZPL (Zebra Programming Language) | USB, Network (TCP/IP) | Industry standard for retail labels |
| DYMO LabelWriter | DYMO XML (via DYMO SDK) | USB | Desktop label printing |
| Brother QL Series | Brother ESC/P | USB, Network | Versatile label printing |
| Receipt Printer (fallback) | ESC/POS | USB, Network | Print labels on 80mm receipt paper |
3.10.3 Batch Printing Workflow
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI / Admin
participant API as Backend
participant PS as Print Service
participant PR as Label Printer
Note over U, PR: Batch Label Printing Workflow
U->>UI: Select products (search, category, or bulk select)
UI-->>U: Display selected products (count, names)
U->>UI: Click "Print Labels"
UI-->>U: Show template selection dialog
U->>UI: Choose label template (e.g., "Standard Barcode 50x25")
UI-->>U: Show quantity options
U->>UI: Set quantity per product
Note right of UI: Default: 1 per product
Note right of UI: Option: "Match stock qty" auto-fills
U->>UI: Click "Preview"
UI->>API: POST /labels/preview
API-->>UI: Return rendered label previews (first 5)
UI-->>U: Display label preview grid
U->>UI: Select target printer from configured list
U->>UI: Click "Print"
UI->>API: POST /labels/print
API->>API: Generate label data for all products
API->>PS: Send print job (template + product data)
PS->>PS: Render labels in printer language (ZPL/DYMO/ESC)
PS->>PR: Send to printer
PR-->>PS: Print confirmation
PS-->>API: Job complete (labels_printed count)
API->>API: Log print job to print_log
API-->>UI: "Printed 45 labels on Zebra-Stockroom"
UI-->>U: Success confirmation
Print Job Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
template_id | UUID | Yes | Label template used |
printer_name | String(100) | Yes | Target printer identifier |
product_count | Integer | Yes | Number of distinct products |
label_count | Integer | Yes | Total labels printed (sum of quantities) |
status | Enum | Yes | QUEUED, PRINTING, COMPLETED, FAILED |
error_message | String | No | Error details if status is FAILED |
initiated_by | UUID | Yes | Staff member who started the print job |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | When the job was created |
completed_at | DateTime | No | When printing finished |
3.10.4 Print Triggers
| Trigger Event | Action | Prompt Behavior | Default Setting |
|---|---|---|---|
| On PO Receive | Prompt to print barcode labels for received items | Modal: “Print labels for 48 received items?” with template selection | Enabled (configurable per tenant) |
| On Transfer Receive | Prompt to print barcode labels for transferred items | Same modal as PO receive | Enabled |
| On Price Change | Prompt to print new shelf tags for price-changed products | Modal: “Price changed for 3 products. Print new shelf tags?” | Enabled |
| On Markdown | Prompt to print clearance stickers for marked-down items | Modal: “Print clearance stickers for 12 items?” with clearance template | Enabled |
| On New Product Published | Prompt to print barcode labels for newly published products | Modal: “Product published. Print labels?” | Disabled (configurable) |
| Manual | Staff initiates from product detail, category view, or bulk selection | No prompt – direct template selection | Always available |
Business Rules:
- Print triggers are configurable per tenant (enable/disable each trigger)
- Staff can dismiss any auto-prompt without printing
- All print events are logged regardless of trigger type
- Print triggers fire only at the location where the event occurred
3.10.5 Reports: Label Printing
| Report | Purpose | Key Data Fields |
|---|---|---|
| Label Print Log | Track all label printing activity | Template used, product count, labels printed, printer, staff member, timestamp, trigger type |
| Reprint Needed | Products with price changes since last label print | Product, current price, price at last label print, last print date, price delta |
| Printer Usage Report | Monitor printer workload and failures | Printer name, jobs processed, labels printed, failure count, avg job size |
| Label Cost Estimate | Estimate label media consumption | Label type, labels printed (period), estimated media cost (based on configured cost per label) |
3.11 Product Media
Scope: Managing product images and video content for POS display, catalog browsing, and Shopify synchronization. Each product supports multiple images with one designated primary, plus optional video links. Images are optimized for fast POS terminal rendering.
3.11.1 Image Management
Product Image Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | Parent product reference |
variant_id | UUID | No | Variant reference (null = product-level image) |
url | String(500) | Yes | Full-resolution image URL |
thumbnail_urls | JSON | Yes | Generated thumbnails: {"sm": "url_64px", "md": "url_128px", "lg": "url_256px"} |
alt_text | String(255) | No | Accessibility description |
sort_order | Integer | Yes | Display position in gallery (1-based) |
is_primary | Boolean | Yes | Primary display image (one per product, one per variant) |
file_size_bytes | Integer | Yes | Original file size for storage tracking |
width_px | Integer | Yes | Original image width |
height_px | Integer | Yes | Original image height |
uploaded_by | UUID | Yes | Staff member who uploaded |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Upload timestamp |
Image Rules:
| Rule | Specification |
|---|---|
| Primary image | One per product; one per variant (optional). Displayed in POS search results, cart line items, and product detail. |
| Gallery images | Up to 20 additional images per product (max configurable per tenant, default 20) |
| Per-variant images | Variants can have their own images (e.g., different image per color). If no variant image, falls back to parent product primary image. |
| Supported formats | JPEG, PNG, WebP |
| Maximum file size | 5MB per image |
| Minimum resolution | 256 x 256 pixels |
| Maximum resolution | 4096 x 4096 pixels (larger images auto-resized on upload) |
| Drag-drop reordering | Staff reorders gallery images by dragging; sort_order updated in batch |
Thumbnail Generation:
On upload, the system auto-generates three thumbnail sizes:
| Size Key | Dimensions | Use Case |
|---|---|---|
sm | 64 x 64 px | POS cart line item, compact list view |
md | 128 x 128 px | POS search results grid |
lg | 256 x 256 px | POS product detail, category browse tile |
Thumbnails are generated asynchronously via a background job. Original image is stored as-is; thumbnails are derived copies.
3.11.2 Video Support
Product Video Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | Parent product reference |
video_url | String(500) | Yes | Video URL (YouTube, Vimeo, or self-hosted) |
video_provider | Enum | No | YOUTUBE, VIMEO, SELF_HOSTED, OTHER |
title | String(255) | Yes | Video display title |
description | String(1000) | No | Brief description of video content |
thumbnail_url | String(500) | No | Custom thumbnail (auto-fetched from provider if not set) |
sort_order | Integer | Yes | Display position |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Creation timestamp |
Video Rules:
- URL-based video links only (no direct video file upload)
- Use cases: product demos, styling guides, assembly instructions, care guides
- Display: video tab on product detail page in admin portal
- Not displayed at POS terminal (bandwidth and performance consideration)
- Maximum 5 videos per product
3.11.3 Media Sync
Shopify Integration:
| Direction | Behavior |
|---|---|
| POS to Shopify (Initial Publish) | Primary image pushed to Shopify on first product publish. Sets as Shopify product featured image. |
| POS to Shopify (Updates) | Primary image updates push to Shopify. Additional gallery images are NOT auto-synced (managed in Shopify separately). |
| Shopify to POS | Not synced. Shopify-managed images remain in Shopify only. POS images are POS-authoritative. |
Image Optimization Pipeline:
sequenceDiagram
autonumber
participant U as Staff
participant UI as Admin UI
participant API as Backend
participant IMG as Image Service
participant CDN as CDN
participant DB as DB
U->>UI: Upload product image (drag-drop or file select)
UI->>UI: Client-side validation (format, size < 5MB)
UI->>API: POST /products/{id}/images (multipart upload)
API->>IMG: Process image
IMG->>IMG: Validate dimensions (min 256x256)
IMG->>IMG: Auto-resize if > 4096x4096
IMG->>IMG: Compress (quality 85%, strip EXIF metadata)
IMG->>IMG: Convert to WebP (if not already)
par Thumbnail Generation
IMG->>IMG: Generate 64x64 thumbnail (sm)
IMG->>IMG: Generate 128x128 thumbnail (md)
IMG->>IMG: Generate 256x256 thumbnail (lg)
end
IMG->>CDN: Upload original + 3 thumbnails
CDN-->>IMG: Return CDN URLs
IMG-->>API: Return URLs (original + thumbnails)
API->>DB: Save image record with URLs
API-->>UI: Image uploaded successfully
UI-->>U: Display new image in gallery
CDN Delivery:
- All images served via CDN URL for fast display across all locations
- CDN cache TTL: 30 days (images are immutable; new upload = new URL)
- Fallback: if CDN is unavailable, images served from origin storage
3.12 Product Notes & Attachments
Scope: Supporting structured internal notes and file attachments on products for buying decisions, vendor communication, quality tracking, and staff communication. Notes and attachments are visible only in the admin portal, not at the POS terminal.
3.12.1 Structured Note Types
| Note Type | Purpose | Typical Authors | Icon |
|---|---|---|---|
| Buying Note | Purchasing decisions, reorder plans, vendor negotiation details | Buyers, Managers | Shopping cart icon |
| Vendor Note | Vendor communication, lead time updates, quality issues, terms changes | Buyers, Admin | Truck icon |
| Quality Note | Quality inspections, defect reports, customer complaints about product | Staff, Managers | Checkmark/shield icon |
| Staff Note | General internal communication about the product | Any staff | Chat bubble icon |
Product Note Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | Parent product reference |
note_type | Enum | Yes | BUYING, VENDOR, QUALITY, STAFF |
content | Text | Yes | Note text (max 5,000 characters) |
is_pinned | Boolean | Yes | Pinned notes display at top (default: false) |
created_by | UUID | Yes | Staff member who created the note |
updated_by | UUID | No | Staff member who last edited (null if never edited) |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Only the note creator or an Admin can edit or delete a note
- Pinned notes always sort to the top, regardless of date
- Maximum 1 pinned note per note type per product (pinning a new note of the same type unpins the previous)
- Notes are never hard-deleted; they are soft-deleted with a
deleted_attimestamp for audit
3.12.2 File Attachments
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | Parent product reference |
file_name | String(255) | Yes | Original uploaded file name |
file_url | String(500) | Yes | Storage URL for download |
file_type | Enum | Yes | PDF, JPEG, PNG, XLSX, DOCX, CSV |
file_size_bytes | Integer | Yes | File size for storage tracking |
description | String(500) | No | Brief description of the attachment |
uploaded_by | UUID | Yes | Staff member who uploaded |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Upload timestamp |
Attachment Rules:
| Rule | Specification |
|---|---|
| Supported file types | PDF, JPEG, PNG, XLSX, DOCX, CSV |
| Maximum file size | 10MB per attachment |
| Maximum attachments per product | 20 |
| Common use cases | Spec sheets, certificates of authenticity, vendor catalogs, care instructions, import documents |
| Access control | Any staff can view; Buyer, Manager, or Admin can upload/delete |
3.12.3 Display
Admin Product Detail Page – Notes Tab:
┌─────────────────────────────────────────────────────────┐
│ Notes (7) [+ Add Note ▼] │
│ ─────────────────────────────────────────────────────── │
│ Filter: [All Types ▼] [Show Pinned Only ☐] │
│ │
│ 📌 BUYING NOTE — Jan 15, 2026 by Sarah (Buyer) │
│ ┌──────────────────────────────────────────────────┐ │
│ │ Reorder 200 units for Spring. Vendor confirmed │ │
│ │ lead time of 21 days. Negotiate 5% volume disc. │ │
│ └──────────────────────────────────────────────────┘ │
│ │
│ QUALITY NOTE — Jan 12, 2026 by Mike (Manager) │
│ ┌──────────────────────────────────────────────────┐ │
│ │ Customer complaint: stitching loose on collar. │ │
│ │ Inspected 5 units from last batch - 2 defective. │ │
│ └──────────────────────────────────────────────────┘ │
│ │
│ STAFF NOTE — Jan 10, 2026 by Jane (Staff) │
│ ┌──────────────────────────────────────────────────┐ │
│ │ This item sells best when displayed near the │ │
│ │ front entrance. Move to endcap for weekends. │ │
│ └──────────────────────────────────────────────────┘ │
│ │
│ ─────────────────────────────────────────────────────── │
│ Attachments (3) [+ Upload File] │
│ │
│ 📎 Nike-SS26-Spec-Sheet.pdf (1.2 MB) [Download] │
│ 📎 Care-Instructions-EN.pdf (340 KB) [Download] │
│ 📎 Vendor-Quote-Jan2026.xlsx (85 KB) [Download] │
└─────────────────────────────────────────────────────────┘
Product List View:
- Note count badge displayed on product rows (e.g., “3 notes”) as a small indicator
- Badge color: grey for staff notes only, yellow if any buying/vendor notes exist, red if any quality notes exist
3.13 Catalog Permissions & Approvals
Scope: Controlling access to catalog features through role-based permissions, field-level edit restrictions, and approval workflows for sensitive changes (pricing, cost, lifecycle transitions). All permission-governed actions are logged to the audit trail defined in Section 3.13.4.
3.13.1 Role-Based Catalog Access
| Permission | Admin | Buyer | Manager | Staff |
|---|---|---|---|---|
| View Products | Yes | Yes | Yes | Yes |
| Create Products | Yes | Yes | Yes | No |
| Edit Products | Yes | Yes | Yes | No (view only) |
| Change Price | Yes | No | Yes | No |
| Change Cost | Yes | Yes | No | No |
| Change Lifecycle Status | Yes | No | Yes | No |
| Delete Products | Yes | No | No | No |
| Approve Changes | Yes | No | Yes | No |
| Manage Categories | Yes | No | Yes | No |
| Manage Vendors | Yes | Yes | No | No |
| Create Purchase Orders | Yes | Yes | Yes | No |
| Submit Purchase Orders | Yes | Yes | No | No |
| Receive Inventory | Yes | Yes | Yes | No |
| Configure Templates/Buttons | Yes | No | Yes | No |
| Export Catalog Data | Yes | Yes | Yes | No |
| Import Catalog Data | Yes | Yes | No | No |
Role Assignment:
- Roles are assigned per user per tenant (a user can have different roles in different tenants)
- A user can hold exactly one catalog role per tenant
- Role assignment requires Admin permission
- Role changes take effect on the user’s next login (or session refresh)
3.13.2 Field-Level Permissions
Configurable per role per tenant. The tenant Admin can customize which fields each role can edit versus view as read-only.
Default Field-Level Restrictions:
| Field(s) | Admin | Buyer | Manager | Staff |
|---|---|---|---|---|
base_price, compare_at_price | Editable | Read-only | Editable | Read-only |
cost, vendor_cost | Editable | Editable | Read-only | Read-only |
lifecycle_status | Editable | Read-only | Editable | Read-only |
category_id | Editable | Editable | Editable | Read-only |
name, description, tags | Editable | Editable | Editable | Read-only |
barcode, alternate_barcodes | Editable | Editable | Editable | Read-only |
images, media | Editable | Editable | Editable | Read-only |
tax_code | Editable | Read-only | Editable | Read-only |
track_inventory, allow_negative | Editable | Read-only | Editable | Read-only |
low_stock_threshold | Editable | Editable | Editable | Read-only |
| Custom attributes | Editable | Editable | Editable | Read-only |
Field Permission Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
role_id | UUID | Yes | Reference to role |
field_name | String(100) | Yes | Product field name (e.g., “base_price”, “cost”) |
permission | Enum | Yes | READ_ONLY or EDITABLE |
tenant_id | UUID | Yes | Tenant scope |
updated_at | DateTime | Yes | Last modification timestamp |
Constraint: Unique on (role_id, field_name, tenant_id).
3.13.3 Approval Workflows
Configurable approval rules determine which catalog changes require manager or admin sign-off before taking effect.
Default Approval Rules:
| Change Type | Condition | Approval Required From | Priority |
|---|---|---|---|
| Price decrease | > 10% reduction from current price | Manager | Medium |
| Price decrease | > 30% reduction from current price | Admin | High |
| Price increase | > 50% increase from current price | Manager | Medium |
| Cost change | Any modification to cost or vendor_cost | Buyer or Admin | Medium |
| Product activation | Draft to Active transition | Manager | Low |
| Product deactivation | Active to Discontinued transition | Manager | Medium |
| Product deletion | Any deletion of Active or Discontinued product | Admin | High |
| Bulk price change | Any bulk operation affecting base_price | Manager | High |
| Bulk status change | Any bulk operation affecting lifecycle_status | Manager | High |
Approval Rule Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
change_type | Enum | Yes | PRICE_DECREASE, PRICE_INCREASE, COST_CHANGE, STATUS_CHANGE, PRODUCT_DELETION, BULK_PRICE, BULK_STATUS |
condition | JSON | Yes | Threshold definition: {"field": "base_price", "operator": "decrease_pct_gt", "value": 10} |
required_role | Enum | Yes | Minimum role required to approve: MANAGER, ADMIN, BUYER_OR_ADMIN |
is_active | Boolean | Yes | Whether this rule is currently enforced |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Approval Request Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
approval_rule_id | UUID | Yes | Rule that triggered this approval |
change_type | Enum | Yes | Type of change requested |
product_id | UUID | Yes | Product affected |
field_name | String(100) | Yes | Field being changed |
old_value | String(500) | Yes | Current value (serialized) |
new_value | String(500) | Yes | Proposed value (serialized) |
reason | String(1000) | No | Requester’s justification |
requested_by | UUID | Yes | Staff member requesting the change |
approved_by | UUID | No | Staff member who approved (null if pending or rejected) |
status | Enum | Yes | PENDING, APPROVED, REJECTED, EXPIRED |
rejection_reason | String(1000) | No | Why the change was rejected |
expires_at | DateTime | No | Auto-expire if not acted on (default: 7 days) |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | When the request was created |
resolved_at | DateTime | No | When the request was approved or rejected |
Approval Workflow:
sequenceDiagram
autonumber
participant U as Staff / Buyer
participant UI as Admin UI
participant API as Backend
participant DB as DB
participant N as Notification Service
participant A as Approver (Manager/Admin)
U->>UI: Edit product field (e.g., reduce price by 25%)
UI->>API: PUT /products/{id}
API->>API: Check approval rules for this change
alt Approval Required
API->>DB: Create approval_request (status: PENDING)
API->>DB: Store proposed change (do NOT apply yet)
API->>N: Send notification to eligible approvers
N-->>A: "Price change requires your approval"
API-->>UI: "Change submitted for approval"
UI-->>U: "Pending manager approval. Price unchanged until approved."
Note over A, DB: Approver Reviews
A->>UI: Open "Pending Approvals" queue
UI->>API: GET /approvals?status=PENDING
API-->>UI: List of pending approval requests
A->>UI: Review change details
Note right of UI: Shows: product, field, old value, new value, who requested, reason
alt Approve
A->>UI: Click "Approve"
UI->>API: POST /approvals/{id}/approve
API->>DB: Update approval_request (status: APPROVED, approved_by, resolved_at)
API->>DB: Apply the change to the product
API->>DB: Log to audit_trail (with approval_id reference)
API->>N: Notify requester "Your change was approved"
API-->>UI: "Change approved and applied"
else Reject
A->>UI: Click "Reject" + enter reason
UI->>API: POST /approvals/{id}/reject
API->>DB: Update approval_request (status: REJECTED, rejection_reason, resolved_at)
API->>N: Notify requester "Your change was rejected: [reason]"
API-->>UI: "Change rejected"
end
else No Approval Required
API->>DB: Apply change directly
API->>DB: Log to audit_trail
API-->>UI: "Change saved"
end
Business Rules:
- Pending approval requests expire after 7 days (configurable per tenant) and auto-set to
EXPIRED - A requester cannot approve their own change request
- If the product is edited again while a pending approval exists for the same field, the older request is auto-cancelled
- Admin can bypass approval requirements (self-approving)
- Approval rules can be enabled or disabled per tenant without deleting the rule definition
3.13.4 Audit Trail
Every field-level change to any catalog product is logged to an immutable audit trail.
Audit Trail Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | Product that was changed |
field_name | String(100) | Yes | Field that changed |
old_value | Text | No | Previous value (null for new records) |
new_value | Text | Yes | New value |
changed_by | UUID | Yes | User who made the change |
change_source | Enum | Yes | MANUAL, IMPORT, SYNC, BULK, SYSTEM, APPROVAL |
approval_id | UUID | No | Reference to approval request (if approval was required) |
ip_address | String(45) | No | IP address of the client |
user_agent | String(500) | No | Client user-agent string |
tenant_id | UUID | Yes | Tenant scope |
created_at | DateTime | Yes | Timestamp of the change |
Audit Trail Rules:
| Rule | Specification |
|---|---|
| Immutability | Audit records cannot be edited or deleted by any user, including Admin |
| Retention | Minimum 7 years (configurable per tenant; can be increased, never decreased) |
| Sensitive field highlighting | Changes to base_price, cost, compare_at_price, and lifecycle_status are visually flagged with a warning icon in the UI |
| Searchable | Audit trail is searchable per product, per user, per field, per date range, and per change source |
| Export | Audit trail exportable as CSV for compliance and external audit |
| Bulk change tracking | Bulk operations create one audit entry per product per field changed (not a single aggregate entry) |
Audit Trail View (Admin Product Detail Page):
┌─────────────────────────────────────────────────────────────────┐
│ Change History (42 changes) [Export CSV] [Filter ▼] │
│ ───────────────────────────────────────────────────────────── │
│ │
│ ⚠ Jan 20, 2026 14:32 — Mike (Manager) — APPROVAL │
│ base_price: $29.99 → $22.99 (approved by Sarah) │
│ │
│ Jan 18, 2026 09:15 — Jane (Buyer) — MANUAL │
│ tags: ["summer"] → ["summer", "clearance"] │
│ │
│ ⚠ Jan 15, 2026 11:00 — SYSTEM — SYNC │
│ lifecycle_status: ACTIVE → DISCONTINUED │
│ │
│ Jan 10, 2026 16:45 — Import Bot — IMPORT │
│ cost: $12.50 → $13.00 │
│ │
│ [Load More...] │
└─────────────────────────────────────────────────────────────────┘
3.13.5 Reports: Permissions & Audit
| Report | Purpose | Key Data Fields |
|---|---|---|
| Pending Approvals | Changes currently awaiting manager/admin approval | Product, change type, field, old value, new value, requested by, requested at, days pending, approver required |
| Approval History | Completed approval decisions with turnaround metrics | Product, change type, requested by, approved/rejected by, resolution time (hours), reason |
| Approval SLA Report | Measure approval response times against targets | Avg resolution time, % resolved within 24h, % expired, by approver |
| Change Audit Log | Complete catalog change history (filterable) | Product, field, old value, new value, changed by, change source, approval ID, timestamp |
| Change Volume Report | Aggregate change counts for workload analysis | Date, change count by source (manual/import/sync/bulk), change count by field, top changers |
| Permission Violations | Attempted unauthorized actions that were blocked | User, role, attempted action, product, timestamp, blocking rule |
3.14 Product Performance Analytics
Scope: Providing real-time product performance metrics embedded on the product detail page and a dedicated catalog analytics dashboard. These metrics drive inventory optimization, markdown decisions, and merchandising strategy.
3.14.1 Embedded Product Metrics
Displayed on each product’s detail page in the admin portal:
| Metric | Calculation | Display Format | Color Coding |
|---|---|---|---|
| Sell-Through Rate | (Units Sold / Units Received) x 100 over selected period | Percentage with trend arrow (up/down vs prior period) | Green >= 70%, Yellow 40-69%, Red < 40% |
| Days of Supply | Current Stock / Avg Daily Sales (rolling 30 days) | Integer with unit “days” | Red < 14 days, Yellow 14-30 days, Green > 30 days |
| Gross Margin % | ((Selling Price - Weighted Avg Cost) / Selling Price) x 100 | Percentage | Red < 30%, Yellow 30-50%, Green > 50% |
| Sales Velocity | Units sold per week (rolling 4-week average) | Decimal units/week with sparkline chart (8-week trend) | No color coding; sparkline shows trend |
| Inventory Aging | Days since last receive at each location | Days per location | Green < 60 days, Yellow 60-120 days, Red > 120 days |
| ABC Classification | Revenue-based Pareto analysis (see 3.14.2) | Badge: A / B / C / NEW | A = Green, B = Blue, C = Grey, NEW = Purple |
| Stock Turn Rate | (COGS / Avg Inventory Value) annualized | Decimal turns/year | Red < 2, Yellow 2-4, Green > 4 |
| Revenue (Period) | Sum of (qty_sold x selling_price) in selected date range | Currency with period selector | No color coding |
Metric Display Layout (Admin Product Detail):
┌─────────────────────────────────────────────────────────────┐
│ Performance Metrics Period: [Last 30 Days ▼] │
│ ───────────────────────────────────────────────────────── │
│ │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ Sell-Thru │ │ Days of │ │ Margin │ │ Velocity │ │
│ │ 72% ▲ │ │ Supply │ │ 54.2% │ │ 8.3/wk │ │
│ │ (Green) │ │ 18 days │ │ (Green) │ │ ~~~~~~~~ │ │
│ │ │ │ (Yellow) │ │ │ │ sparkline│ │
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘ │
│ │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │ Aging │ │ ABC │ │ Turns │ │ Revenue │ │
│ │ 45 days │ │ [A] │ │ 6.2/yr │ │ $12,450 │ │
│ │ (Green) │ │ (Green) │ │ (Green) │ │ │ │
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘ │
└─────────────────────────────────────────────────────────────┘
3.14.2 ABC Classification
Calculation Method: Revenue-based Pareto analysis, calculated monthly via background job.
| Class | Revenue Contribution | Product Population | Attention Level |
|---|---|---|---|
| A | Top 20% of products generating ~80% of revenue | ~20% of active SKUs | High: frequent cycle counts, optimal stock levels, priority reorder |
| B | Next 30% of products generating ~15% of revenue | ~30% of active SKUs | Moderate: standard stock levels, regular review |
| C | Bottom 50% of products generating ~5% of revenue | ~50% of active SKUs | Low: review for markdown or discontinuation, minimal safety stock |
| NEW | Products active for < 60 days (insufficient data) | Varies | Exempt from classification until data accumulates |
ABC Classification Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | Product reference |
classification | Enum | Yes | A, B, C, NEW |
revenue_total | Decimal(12,2) | Yes | Revenue in the analysis period |
revenue_rank_pct | Decimal(5,2) | Yes | Percentile rank (0-100) by revenue |
analysis_period_start | Date | Yes | Start of the analysis period |
analysis_period_end | Date | Yes | End of the analysis period |
tenant_id | UUID | Yes | Tenant scope |
calculated_at | DateTime | Yes | When this classification was computed |
Business Rules:
- Recalculated on the 1st of each month using the trailing 12-month sales period
- Products with fewer than 60 days since
published_atare classified asNEW - Classification stored on the product record as
abc_classificationfield for fast query - Classification is tenant-scoped (each tenant’s products are ranked independently)
- ABC drives: reorder priority (A items reordered first), count scheduling (A items counted more frequently), display prominence in analytics
ABC Classification Impact on Operations:
| Area | A Items | B Items | C Items |
|---|---|---|---|
| Cycle Count Frequency | Weekly | Monthly | Quarterly |
| Safety Stock | 2 weeks of supply | 1 week of supply | Minimum (3 days) |
| Reorder Priority | First in auto-PO generation | Second priority | Only if manually flagged |
| Stockout Alerting | Immediate alert to Manager | End-of-day alert | Weekly report only |
| Markdown Review | Quarterly review | Semi-annual review | Monthly review (candidates for clearance) |
3.14.3 Catalog Analytics Dashboard
A dedicated dashboard accessible from the admin portal navigation.
Summary Cards (Top Row):
| Card | Metric | Calculation |
|---|---|---|
| Total Active SKUs | Count of products where lifecycle_status = ACTIVE | Real-time |
| Total Inventory Value | Sum of (qty_on_hand x weighted_avg_cost) across all locations | Refreshed hourly |
| Avg Gross Margin % | Average margin across all active products (weighted by revenue) | Refreshed daily |
| Avg Days of Supply | Average days of supply across all active products | Refreshed daily |
Charts:
| Chart | Type | Data | Purpose |
|---|---|---|---|
| Top 10 by Revenue | Horizontal bar chart | Top 10 products by revenue in selected period | Identify best sellers |
| Bottom 10 by Sell-Through | Horizontal bar chart | Bottom 10 products by sell-through rate | Candidates for markdown or discontinuation |
| ABC Distribution | Pie chart (dual-ring) | Inner ring: product count per class. Outer ring: revenue per class | Visualize catalog health |
| Inventory Aging Histogram | Stacked bar chart | Buckets: 0-30, 30-60, 60-90, 90-120, 120+ days | Identify aging inventory |
| Category Margin Comparison | Horizontal bar chart | Avg margin by top-level category | Compare category profitability |
| Seasonal Sell-Through | Multi-line chart | Sell-through rate per season over time | Compare seasonal performance |
| Velocity vs Stock Scatter | Scatter plot | X = velocity, Y = stock qty, color = ABC class | Spot overstock/understock |
Dashboard Filters:
| Filter | Type | Scope |
|---|---|---|
| Date Range | Date picker | All charts and metrics |
| Location | Multi-select | Filter to specific store(s) or all |
| Category | Tree-select | Filter to specific category branch |
| Brand | Multi-select | Filter to specific brand(s) |
| Season | Multi-select | Filter to season collection |
| ABC Class | Multi-select | Filter by A, B, C, or NEW |
3.14.4 Reports: Analytics
| Report | Purpose | Key Data Fields |
|---|---|---|
| Product Scorecard | Full performance summary per product, exportable | Product, SKU, revenue, units sold, margin %, velocity, aging, ABC class, sell-through %, days of supply, stock turn |
| Slow Movers | Products with low velocity and high stock (markdown candidates) | Product, stock qty, days of supply, velocity, last sale date, ABC class, recommended action |
| Overstock Alert | Products exceeding target days of supply threshold | Product, location, stock qty, target DOS, actual DOS, excess units, excess value at cost |
| Dead Stock | Zero sales in configurable period (default 90 days) | Product, last sale date, stock qty, stock value at cost, days without sale, ABC class |
| Margin Erosion | Products where margin has declined over time | Product, margin 90 days ago, current margin, margin delta, cause (cost increase / price decrease / promo) |
| ABC Migration | Products that changed classification between periods | Product, previous class, current class, revenue change, velocity change |
3.15 Catalog User Stories & Acceptance Criteria (EXPANDED)
Note: Epics 3.A through 3.E and their existing Gherkin scenarios remain unchanged. The following adds new Epics 3.F through 3.S and new Gherkin acceptance criteria for key features.
Epic 3.F: Pricing Engine
- Story 3.F.1 (Price Hierarchy): System resolves product price using cascading hierarchy: Manual Override > Promotion > Price Book > Channel Price > Global Default. The customer always receives the best applicable price.
- Story 3.F.2 (Price Books): Admin can create named price books restricted by customer group, channel, and date range. Price book entries override base price for matched products.
- Story 3.F.3 (Promotions): Staff can create four promotion types: Basic (% or $ off), Tiered (volume pricing), BOGO (cross-item), and Scheduled (time-based). Promotions follow a Draft > Scheduled > Active > Expired lifecycle.
- Story 3.F.4 (Markdown Workflow): Price reductions follow a formal workflow: request > manager approval > scheduled price change. All price changes logged with who/when/why for accountability.
- Story 3.F.5 (Conflict Resolution): When multiple pricing rules match, system applies best-price-for-customer logic. Exclusive promotions override all other pricing. Stackable promotions combine up to configurable max discount.
Epic 3.G: Multi-Channel
- Story 3.G.1 (Channel Visibility): Staff can toggle product visibility per channel (In-Store, Online, Wholesale). Products must be visible on at least one channel to be Active.
- Story 3.G.2 (Channel Inventory): Admin can choose shared pool (default) or dedicated allocation per channel. Dedicated mode reserves specific quantities per channel.
- Story 3.G.3 (Channel Pricing): Products can have different prices per channel. Channel price overrides base price per the price hierarchy.
Epic 3.H: Shopify Integration
- Story 3.H.1 (POS-Master Sync): Product changes in POS automatically push to Shopify for POS-owned fields. Shopify-only fields (SEO, metafields) are preserved.
- Story 3.H.2 (Bidirectional Mode): Admin can enable bidirectional sync per tenant. POS-priority conflict resolution applies to shared fields.
- Story 3.H.3 (Sync Monitoring): Admin dashboard shows sync status, pending changes, failed syncs, and conflict log.
Epic 3.I: Search & Discovery
- Story 3.I.1 (Full-Text Search): POS search supports name, SKU, barcode, tags, vendor, brand, and custom attributes with fuzzy matching and auto-complete. Results return in under 200ms.
- Story 3.I.2 (Favorites & Quick-Add): Staff can pin up to 50 favorite products for one-tap access and configure up to 20 quick-add buttons on the POS home screen.
- Story 3.I.3 (Substitutions): When a product is out of stock, system suggests alternatives: same product at other locations, similar products in same category, and related products from cross-sell configuration.
Epic 3.J: Labels & Printing
- Story 3.J.1 (Label Printing): Staff can select products and print barcode labels, shelf tags, or clearance stickers using configurable templates. Supports Zebra, DYMO, Brother, and receipt printer output.
- Story 3.J.2 (Auto-Print Triggers): System prompts label printing on PO receive, transfer receive, price change, and markdown events. Triggers are configurable per tenant.
Epic 3.K: Media Management
- Story 3.K.1 (Product Images): Products support one primary image plus a gallery of up to 20 images. Per-variant images are supported. Drag-drop reorder. Auto-generated thumbnails (64px, 128px, 256px).
- Story 3.K.2 (Video): Products can link to video URLs (YouTube, Vimeo, self-hosted) for demos and styling guides. Videos displayed on admin product detail page only (not POS terminal).
Epic 3.L: Permissions & Approvals
- Story 3.L.1 (Role-Based Access): Four catalog roles (Admin, Buyer, Manager, Staff) with configurable field-level permissions. Staff is view-only. Admin has full access.
- Story 3.L.2 (Approval Workflows): Price decreases > 10% require Manager approval. Price decreases > 30% require Admin approval. Cost changes require Buyer/Admin approval. Rules are configurable per tenant.
- Story 3.L.3 (Audit Trail): Every field change logged with who, when, old value, new value, and change source. Searchable per product. Minimum 7-year retention. Exportable as CSV.
Epic 3.M: Analytics
- Story 3.M.1 (Product Metrics): Product detail page shows sell-through rate, days of supply, gross margin, sales velocity (sparkline), inventory aging, ABC classification, stock turn rate, and period revenue.
- Story 3.M.2 (ABC Classification): Monthly Pareto analysis classifies products as A (top 20% by revenue), B (next 30%), C (bottom 50%). New products exempt until 60 days of data.
- Story 3.M.3 (Analytics Dashboard): Dedicated catalog dashboard with summary cards, top/bottom performers, ABC distribution chart, aging histogram, category margin comparison, and seasonal sell-through trends.
Catalog Acceptance Criteria: New Gherkin Scenarios
Feature: Price Hierarchy Resolution
Feature: Price Hierarchy Resolution
As a POS system
I need to resolve the correct price using the pricing hierarchy
So that customers always receive the best applicable price
Background:
Given product "Classic Tee" has base_price "$29.99"
And product "Classic Tee" is in "ACTIVE" lifecycle status
Scenario: Base price used when no overrides exist
Given no channel price, price book, or promotion applies to "Classic Tee"
When any customer adds "Classic Tee" to cart
Then the price should be "$29.99"
Scenario: Channel price overrides base price
Given channel "WHOLESALE" has price "$19.99" for "Classic Tee"
When a wholesale customer adds "Classic Tee" to cart
Then the price should be "$19.99"
Scenario: Price book overrides channel price
Given price book "Employee Discount" is active for customer group "Employees"
And "Employee Discount" has "Classic Tee" at "$14.99"
And channel "IN_STORE" has price "$29.99" for "Classic Tee"
When employee "John" adds "Classic Tee" to cart
Then the price should be "$14.99"
Scenario: Promotion overrides price book when better for customer
Given promotion "Summer Sale" is active with 30% off "Classic Tee"
And price book "Employee Discount" has "Classic Tee" at "$14.99"
When employee "John" adds "Classic Tee" to cart
Then the price should be "$14.99"
And the system should apply best-price-for-customer logic
And the applied pricing source should be "PRICE_BOOK"
Scenario: Promotion wins when it gives the lower price
Given promotion "Flash Sale" is active with 60% off "Classic Tee"
And price book "Employee Discount" has "Classic Tee" at "$14.99"
When employee "John" adds "Classic Tee" to cart
Then the price should be "$12.00"
And the applied pricing source should be "PROMOTION"
Scenario: Manual override beats all other pricing
Given promotion "Summer Sale" is active with 30% off "Classic Tee"
And price book "Employee Discount" has "Classic Tee" at "$14.99"
When manager "Mike" applies manual override price "$10.00" to "Classic Tee"
Then the price should be "$10.00"
And the applied pricing source should be "MANUAL_OVERRIDE"
And the override should be logged to the audit trail
Scenario: Exclusive promotion overrides stackable promotions
Given exclusive promotion "VIP Members Only" is active with 40% off "Classic Tee"
And stackable promotion "Summer Sale" is active with 10% off "Classic Tee"
When a VIP customer adds "Classic Tee" to cart
Then only the exclusive promotion should apply
And the price should be "$17.99"
Scenario: Stackable promotions combine up to max discount
Given stackable promotion "Summer Sale" is active with 10% off "Classic Tee"
And stackable promotion "Newsletter Signup" is active with 5% off "Classic Tee"
And tenant max discount is configured at 50%
When a qualifying customer adds "Classic Tee" to cart
Then both promotions should apply
And the combined discount should be 15%
And the price should be "$25.49"
Feature: Markdown Workflow with Approval
Feature: Markdown Workflow with Approval
As a catalog manager
I need price reductions to follow an approval workflow
So that markdowns are controlled and accountable
Background:
Given product "Slow Seller" has base_price "$49.99"
And approval rule exists: price decrease > 10% requires Manager approval
And approval rule exists: price decrease > 30% requires Admin approval
Scenario: Small price decrease requires no approval
When staff "Jane" changes price to "$47.99" (4% decrease)
Then the price should change immediately to "$47.99"
And the change should be logged to the audit trail
And no approval request should be created
Scenario: Moderate price decrease requires manager approval
When staff "Jane" requests markdown to "$39.99" (20% decrease) with reason "Low sell-through"
Then an approval request should be created with status "PENDING"
And the product price should remain "$49.99" until approved
And manager "Mike" should receive a notification
Scenario: Manager approves the markdown
Given a pending approval exists for "Slow Seller" price change to "$39.99"
When manager "Mike" approves the markdown
Then the product price should change to "$39.99"
And the approval status should be "APPROVED"
And the audit log should record the change with requester "Jane", approver "Mike", and reason "Low sell-through"
And staff "Jane" should receive a notification "Your markdown was approved"
Scenario: Manager rejects the markdown
Given a pending approval exists for "Slow Seller" price change to "$39.99"
When manager "Mike" rejects the markdown with reason "Wait for end-of-season clearance"
Then the product price should remain "$49.99"
And the approval status should be "REJECTED"
And staff "Jane" should receive a notification with the rejection reason
Scenario: Large price decrease escalates to admin
When staff "Jane" requests markdown to "$29.99" (40% decrease) with reason "Clearance"
Then an approval request should be created requiring Admin approval
And manager approval should NOT be sufficient
Scenario: Approval request expires after 7 days
Given a pending approval was created 8 days ago for "Slow Seller"
When the expiration job runs
Then the approval status should change to "EXPIRED"
And the product price should remain unchanged
And the requester should be notified of expiration
Scenario: Requester cannot approve their own change
When staff "Jane" requests markdown to "$39.99"
And "Jane" also has Manager role
Then "Jane" should not be able to approve her own request
And the system should show "Cannot approve your own change request"
Feature: POS-Shopify Catalog Sync
Feature: POS-Shopify Catalog Sync
As a multi-channel retailer
I need product changes to sync between POS and Shopify
So that online and in-store catalogs stay consistent
Background:
Given product "Classic Tee" exists in POS with SKU "BLK-TEE-001"
And product "Classic Tee" is synced with Shopify product ID "shop_12345"
Scenario: POS price change pushes to Shopify in POS-Master mode
Given sync mode is "POS-Master"
When staff changes price from "$29.99" to "$24.99" in POS
Then the price should update in Shopify within the sync interval
And Shopify SEO title should remain unchanged
And Shopify metafields should remain unchanged
And sync log should record: field "price", direction "POS_TO_SHOPIFY", status "SUCCESS"
Scenario: POS-owned fields are protected from Shopify overwrite
Given sync mode is "POS-Master"
When someone edits the price to "$27.99" directly in Shopify
Then the next sync cycle should overwrite Shopify price back to "$24.99"
And a conflict entry should be logged with source "SHOPIFY", rejected value "$27.99"
Scenario: Shopify description change syncs in bidirectional mode
Given sync mode is "Bidirectional with POS Priority"
And "long_description" is configured with sync direction "Configurable"
And "long_description" is set to "Shopify-to-POS" direction
When an SEO agency updates the description in Shopify
Then the new description should sync to POS on the next cycle
And POS-owned fields (price, SKU, variants) should not be affected
And sync log should record: field "long_description", direction "SHOPIFY_TO_POS", status "SUCCESS"
Scenario: Conflict resolution when both sides change the same field
Given sync mode is "Bidirectional with POS Priority"
And "base_price" is a POS-priority field
When POS changes price to "$24.99" between sync cycles
And Shopify changes price to "$27.99" between the same sync cycles
Then the POS price "$24.99" should win
And Shopify should be updated to "$24.99"
And a conflict audit entry should be created
And the conflict log should show: POS value "$24.99" (applied), Shopify value "$27.99" (rejected)
Scenario: New product publish creates Shopify listing
Given product "New Arrival Shirt" is in "DRAFT" status in POS
And the product has a primary image
When staff publishes the product (Draft to Active)
And channel "Online" is enabled for this product
Then a new Shopify product should be created
And the primary image should be pushed to Shopify
And Shopify product status should be set to "active"
And the Shopify product ID should be stored in the sync mapping table
Scenario: Sync failure is logged and retried
Given sync mode is "POS-Master"
When staff changes price in POS
And the Shopify API returns a 500 error
Then the sync should be marked as "FAILED" in the sync log
And the change should be queued for retry
And after 3 failed retries, the sync should be flagged for manual review
And the admin sync dashboard should show the failure
Feature: Label Printing
Feature: Label Printing
As a retail staff member
I need to print barcode labels and price tags
So that products are properly tagged for sale
Scenario: Print barcode labels for received PO items
Given PO "PO-2026-00042" has just been received with 48 units of "Air Max 90"
When the receive is confirmed
Then the system should prompt "Print labels for 48 received items?"
When staff clicks "Print Labels"
And selects template "Standard Barcode 50x25"
And sets quantity to 48
And selects printer "Zebra-Stockroom"
And clicks "Print"
Then 48 labels should be sent to "Zebra-Stockroom"
And the print log should record the job
Scenario: Auto-prompt on price change
Given product "Classic Tee" has a printed shelf tag at "$29.99"
When manager changes price to "$24.99"
Then the system should prompt "Price changed. Print new shelf tags?"
And the "Reprint Needed" report should include "Classic Tee"
Scenario: Batch print clearance stickers
Given 12 products have been marked down for clearance
When staff selects all 12 products from the clearance collection
And clicks "Print Labels"
And selects template "Clearance Sticker 40x30"
Then each label should show markdown price, original price (strikethrough), discount %, and "CLEARANCE" badge
And 12 labels should be printed
Scenario: Print labels on receipt printer fallback
Given no dedicated label printer is configured at "Store C"
When staff attempts to print labels
Then the system should offer the receipt printer as fallback
And labels should be formatted for 80mm receipt paper width
Feature: Product Search at POS
Feature: Product Search at POS
As a POS operator
I need to find products quickly using multiple search methods
So that checkout is fast and efficient
Scenario: Fuzzy search handles typos
Given product "Oxford Button-Down Shirt" exists
When staff types "oxfrd" in the search bar
Then "Oxford Button-Down Shirt" should appear in search results
And results should return within 200ms
Scenario: SKU exact match ranks highest
Given product "Classic Tee" has SKU "BLK-TEE-001"
And product "Black Tee Dress" has name containing "Tee"
When staff searches "BLK-TEE-001"
Then "Classic Tee" should be the first result (exact SKU match)
Scenario: Auto-complete shows suggestions after 2 characters
When staff types "Cl" in the search bar
Then auto-complete should show suggestions including "Classic Tee", "Classic Oxford", "Clearance Items"
And suggestions should appear within 150ms
Scenario: Out-of-stock product shows substitution suggestions
Given product "Classic Tee" has 0 units available at current location
And "Store B" has 12 units of "Classic Tee"
And "V-Neck Tee" is in the same category at "$27.99"
When staff searches and selects "Classic Tee"
Then the system should show "Out of stock at this location"
And suggest: "Available at Store B (12 units)"
And suggest: "Similar: V-Neck Tee $27.99 (8 in stock)"
Scenario: Quick-add button adds product to cart in one tap
Given "Classic Tee" is configured as quick-add button at position 1
When staff taps the quick-add button
Then "Classic Tee" should be added to cart with qty 1
And no search or product detail view should be required
Scenario: Saved filter retrieves matching products
Given staff saved a filter named "Nike Low Stock" with brand "Nike" and stock status "Low Stock"
When staff selects "Nike Low Stock" from saved filters
Then only Nike products with stock below low_stock_threshold should be displayed
Feature: Catalog Permissions
Feature: Catalog Role-Based Permissions
As a tenant admin
I need to control who can edit catalog fields
So that sensitive data is protected from unauthorized changes
Scenario: Staff role is view-only
Given user "Tom" has role "Staff"
When "Tom" opens product "Classic Tee" detail page
Then all fields should be displayed as read-only
And no "Edit" or "Save" buttons should be visible
And "Tom" should not see "Create Product" option in navigation
Scenario: Buyer cannot change price
Given user "Sarah" has role "Buyer"
When "Sarah" edits product "Classic Tee"
Then the "base_price" field should be read-only
And the "compare_at_price" field should be read-only
But the "cost" field should be editable
And the "vendor_cost" field should be editable
Scenario: Manager cannot change cost
Given user "Mike" has role "Manager"
When "Mike" edits product "Classic Tee"
Then the "cost" field should be read-only
And the "vendor_cost" field should be read-only
But the "base_price" field should be editable
And the "lifecycle_status" field should be editable
Scenario: Unauthorized action is blocked and logged
Given user "Tom" has role "Staff"
When "Tom" attempts to call PUT /products/{id} via API
Then the request should return 403 Forbidden
And a permission violation should be logged with user, role, and attempted action
And the violation should appear in the "Permission Violations" report
Feature: Product Analytics
Feature: Product Performance Analytics
As a merchandising manager
I need to see product performance metrics
So that I can make data-driven inventory and pricing decisions
Scenario: Product detail shows embedded metrics
Given product "Classic Tee" has been active for 90 days
And has sold 450 units out of 600 received
And current stock is 150 units across all locations
And avg daily sales is 5 units
When manager opens the product detail page
Then sell-through rate should show "75%"
And days of supply should show "30 days"
And sales velocity should show "35.0/wk" with sparkline
And ABC classification badge should be displayed
Scenario: ABC classification calculated monthly
Given it is the 1st of the month
And the monthly ABC job runs
Then the top 20% of products by trailing 12-month revenue should be classified "A"
And the next 30% should be classified "B"
And the bottom 50% should be classified "C"
And products active less than 60 days should be classified "NEW"
Scenario: Dead stock report identifies zero-sale products
Given product "Forgotten Widget" has had zero sales for 95 days
And product "Forgotten Widget" has 30 units on hand
And the dead stock threshold is configured at 90 days
When manager runs the "Dead Stock" report
Then "Forgotten Widget" should appear with last_sale_date, stock_qty 30, and days_without_sale 95
And recommended action should be "Review for markdown or discontinuation"
Scenario: Overstock alert fires for excess inventory
Given product "Winter Coat" has target days of supply of 30
And "Store A" has 200 units and sells 1/day (200 days of supply)
When the overstock alert report runs
Then "Winter Coat" at "Store A" should be flagged
And excess units should show 170 (200 - 30 days x 1/day)
And excess value should show the cost of 170 units
4. Inventory Module
Module 4: Inventory Management (Sections 4.1 – 4.7)
4.1 Overview & Scope
The Inventory Management module governs the complete lifecycle of physical stock within the multi-tenant POS system – from procurement through vendor purchase orders, to warehouse and store receiving, through internal logistics and transfers, and into auditing via physical counts and manual adjustments. It is the operational backbone that ensures every unit of merchandise is tracked, accounted for, and available for sale at the right location at the right time.
4.1.1 Executive Summary
Retail clothing operations across five stores and one HQ warehouse demand real-time, accurate inventory visibility. A single garment may be ordered from a vendor, received at the warehouse, transferred to a retail store, reserved for a customer’s online order, counted during a cycle count, adjusted after discovery of damage, and ultimately sold at the point of sale. Each of these events must be captured, validated, and reflected in the system’s inventory balances within seconds.
Module 4 provides the business rules, workflows, data models, and integration points that make this possible. It covers seven functional domains:
- Procurement – Creating, approving, submitting, and tracking purchase orders to vendors (Section 4.3).
- Receiving – Inspecting and accepting inbound inventory from any source – PO shipments, transfers, customer returns, and vendor RMA replacements (Section 4.4).
- Logistics – Inter-store and warehouse-to-store transfers are documented in Module 5 (Transfers & Logistics). Module 4 provides the inventory status and reservation primitives that Module 5 depends on.
- Auditing – Physical stock counts and manual adjustments that reconcile system quantities with reality (Sections 4.6 and 4.7).
- Costing – Inventory valuation via weighted average cost is applied at receiving time and propagated through the system. Cost data feeds into Module 1 (Sales) for margin calculation.
- Integration – Inventory events trigger real-time updates to the POS terminals (Module 1), the catalog (Module 3), and the movement history audit trail.
- Operations – Reorder management, dead stock detection, and minimum display quantity monitoring ensure proactive inventory health (Section 4.5).
4.1.2 Module Dependencies
Module 4 does not operate in isolation. It depends on and is depended upon by multiple other modules in the system.
flowchart LR
M1["Module 1\nSales & POS"]
M3["Module 3\nCatalog"]
M4["Module 4\nInventory"]
M5["Module 5\nTransfers & Logistics"]
M6["Module 6\nReporting"]
M3 -->|Product data, variants,\nvendor links, barcodes| M4
M4 -->|Available qty per location,\nreservation status| M1
M1 -->|Sale committed → decrement,\nVoid → release reservation| M4
M4 -->|Inventory status,\nreservation holds| M5
M5 -->|Transfer receive → increment,\nTransfer ship → decrement| M4
M4 -->|Stock levels, velocity,\ncost data| M6
M1 -->|Sales velocity data| M4
style M4 fill:#2d6a4f,stroke:#1b4332,color:#fff
style M1 fill:#264653,stroke:#1d3557,color:#fff
style M3 fill:#264653,stroke:#1d3557,color:#fff
style M5 fill:#264653,stroke:#1d3557,color:#fff
style M6 fill:#264653,stroke:#1d3557,color:#fff
Upstream dependencies (Module 4 consumes):
| Source Module | Data Consumed | Purpose |
|---|---|---|
| Module 3 (Catalog) | Product ID, variant ID, SKU, barcode, vendor-product links, vendor cost | Identify what is being counted, received, or ordered. Vendor cost used for PO line items. |
| Module 1 (Sales) | Sales velocity per product per location, sale events (commit, void, cancel) | Drive reorder point calculations and inventory decrements/releases. |
| Module 5 (Transfers) | Transfer ship and receive events | Trigger IN_TRANSIT status changes and inventory increments at destination. |
Downstream consumers (Module 4 provides):
| Consumer Module | Data Provided | Purpose |
|---|---|---|
| Module 1 (Sales) | Available quantity per product per location, reservation status | POS checks available qty before completing a sale. Displays stock info to staff. |
| Module 5 (Transfers) | Inventory status, available qty, reservation holds | Transfer system checks available qty before allowing outbound shipment. |
| Module 6 (Reporting) | Stock levels, cost data, velocity, count variances, adjustment history | Inventory reports, shrinkage analysis, days-of-supply calculations. |
4.1.3 Functional Scope
The following table enumerates the functional areas covered by Module 4 and their section references.
| Domain | Section | Description |
|---|---|---|
| Inventory Status Model | 4.2 | Six-status state machine governing what can be sold, transferred, or must be held. |
| Reservation Model | 4.2 | Reserve inventory for carts, parked transactions, transfers, online orders, and hold-for-pickup. |
| Minimum Display Quantity | 4.2 | Advisory warnings when stock drops below configured floor display minimums. |
| Purchase Orders | 4.3 | Full PO lifecycle from draft through receiving and close. |
| PO Approval Workflow | 4.3 | Threshold-based approval routing for high-value purchase orders. |
| Receiving & Inspection | 4.4 | Unified receiving workflow for POs, transfers, returns, and RMA replacements. |
| Discrepancy Handling | 4.4 | Triple-approach handling of receiving variances: note, RMA draft, quarantine. |
| Non-PO Receiving | 4.4 | Accept inventory without a purchase order using mandatory reason codes. |
| Return-to-Stock | 4.4 | Customer return items re-enter available inventory. |
| Reorder Management | 4.5 | Velocity-based reorder points with auto-generated draft POs. |
| Static Override | 4.5 | Manager-locked manual reorder points overriding dynamic calculations. |
| Dead Stock Detection | 4.5 | Alert on products with zero sales velocity over configurable period. |
| Inventory Counting | 4.6 | Five count types with configurable freeze/snapshot modes. |
| Inventory Adjustments | 4.7 | Manual corrections with mandatory manager approval and custom reason codes. |
4.1.4 Key Business Rules Summary
The following rules apply across all Module 4 operations:
- Only AVAILABLE stock can be sold at POS. Inventory in any other status (QUARANTINE, DAMAGED, RESERVED, IN_TRANSIT, PENDING_INSPECTION) is excluded from the sellable quantity displayed to cashiers.
- Only AVAILABLE stock can be transferred between locations. Transfer requests that would reduce available stock below zero are rejected.
- Every inventory change is logged. All status transitions, quantity changes, and cost updates create movement records in the audit trail (see Module 3, Section 3.16 – Movement History).
- Inventory is tracked per product, per variant, per location, per status. A single product may have quantities spread across multiple statuses at a single location simultaneously.
- All monetary values use the tenant’s configured currency. Multi-currency is not supported in v1.
- Tenant isolation is enforced at the data layer. Every inventory record carries a
tenant_idforeign key. Cross-tenant queries are impossible by design.
4.1.5 Inventory Balance Equation
The system maintains the following balance equation at all times for each product-variant-location combination:
Available = On-Hand - Reserved - In-Transit - Quarantine - Damaged
Where:
| Term | Definition |
|---|---|
| On-Hand | Total physical units at the location (all statuses combined). This is what you would count if you physically counted every item. |
| Available | Units that can be sold or transferred right now. This is the number displayed to POS staff. |
| Reserved | Units allocated to a pending sale cart, parked transaction, outbound transfer, online order, or hold-for-pickup. Not yet physically moved but committed. |
| In-Transit | Units that have shipped from this location to another location but have not yet been received at the destination. Decremented from source available, not yet incremented at destination. |
| Quarantine | Units held for inspection or quality review. Cannot be sold or transferred. |
| Damaged | Units identified as damaged. Cannot be sold. May be written off or returned to vendor via RMA. |
The system does not store Available as a separate field. It is always computed from the status-based quantity fields. This ensures the balance equation is always consistent and cannot drift due to bugs in update logic.
4.2 Inventory Status Model
4.2.1 Inventory Status State Machine
Each unit of inventory at each location carries a status that controls whether it can be sold, transferred, or must be held for inspection. The system supports six statuses organized into a state machine with well-defined transitions.
stateDiagram-v2
[*] --> AVAILABLE: Stock Received & Inspected
AVAILABLE --> QUARANTINE: Quality Concern Flagged
AVAILABLE --> RESERVED: Allocated to Order/Transfer
AVAILABLE --> DAMAGED: Damage Identified
QUARANTINE --> AVAILABLE: Inspection Passed
QUARANTINE --> DAMAGED: Inspection Failed
PENDING_INSPECTION --> AVAILABLE: Inspection Passed
PENDING_INSPECTION --> QUARANTINE: Needs Further Review
PENDING_INSPECTION --> DAMAGED: Inspection Failed
DAMAGED --> WRITE_OFF: Unrepairable
DAMAGED --> VENDOR_RMA: Return to Vendor
IN_TRANSIT --> AVAILABLE: Transfer Received & OK
IN_TRANSIT --> PENDING_INSPECTION: Received - Needs Inspection
RESERVED --> AVAILABLE: Reservation Released
note right of AVAILABLE
Sellable at POS
Transferable between locations
end note
note right of QUARANTINE
Blocked from sale
Blocked from transfer
Awaiting inspection
end note
note right of DAMAGED
Blocked from sale
Can be written off or returned to vendor
end note
note right of RESERVED
Allocated but not yet shipped/sold
Decremented from available count
end note
note right of IN_TRANSIT
Moving between locations
Not available at source or destination
end note
Status Definitions:
| Status | Sellable | Transferable | Description |
|---|---|---|---|
AVAILABLE | Yes | Yes | Stock is on the shelf and ready for sale or transfer. |
QUARANTINE | No | No | Stock is held pending quality inspection. Triggered by a staff member flagging a quality concern, or by a receiving inspection that requires further review. |
DAMAGED | No | No | Stock is identified as damaged and cannot be sold. Terminal states from here are WRITE_OFF (removed from inventory) or VENDOR_RMA (returned to vendor for credit or replacement). |
PENDING_INSPECTION | No | No | Stock has arrived (from transfer or PO) and needs inspection before it can be placed on the sales floor. |
RESERVED | No | No | Stock is allocated to a specific purpose (sale cart, parked transaction, outbound transfer, online order, or hold-for-pickup) but has not yet been physically moved or sold. |
IN_TRANSIT | No | No | Stock has shipped from the source location but has not yet arrived at the destination location. It is not available at either location during transit. |
Business Rules:
- Only
AVAILABLEstock can be sold at POS. The POS terminal displays only the AVAILABLE quantity as the sellable count. - Only
AVAILABLEstock can be transferred between locations. Transfer requests that would reduce AVAILABLE stock below zero at the source location are rejected. QUARANTINEandDAMAGEDstock is blocked from sale and transfer. It must be inspected and resolved before it can re-enter the sellable pool.RESERVEDstock is decremented from the available count but not yet physically moved. If the reservation is released (e.g., cart abandoned, parked transaction voided), the stock returns to AVAILABLE.- All status changes require a reason code and are logged to the movement history audit trail.
- Status changes can only follow the transitions defined in the state machine above. Any attempt to make an invalid transition (e.g., QUARANTINE directly to RESERVED) is rejected by the API.
Inventory Status Data Model
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | Reference to product (FK to catalog) |
variant_id | UUID | No | Reference to specific variant, if applicable (FK to catalog) |
location_id | UUID | Yes | Reference to store/warehouse location |
status | Enum | Yes | AVAILABLE, QUARANTINE, DAMAGED, PENDING_INSPECTION, RESERVED, IN_TRANSIT |
qty | Integer | Yes | Quantity in this status at this location |
last_status_change_at | DateTime | Yes | Timestamp of most recent status change |
changed_by | UUID | Yes | User who made the status change |
reason_code | String | Yes | Reason for current status (e.g., QUALITY_CONCERN, TRANSFER_ALLOCATED, TRANSIT_DAMAGE, CART_RESERVE, PARKED_RESERVE) |
tenant_id | UUID | Yes | Owning tenant |
4.2.2 Reservation Model
Reservations temporarily hold inventory for a specific purpose, preventing it from being sold or transferred to another customer or location. The reservation model is central to ensuring that the POS system does not oversell stock in a multi-terminal, multi-channel environment.
When a reservation is created, the specified quantity is moved from AVAILABLE status to RESERVED status. When the reservation is committed (sale completed, transfer shipped), the reserved quantity is decremented from inventory. When the reservation is released (cart abandoned, transaction voided), the reserved quantity returns to AVAILABLE.
Reservation Types
The system supports five distinct reservation types, each with its own lifecycle and rules:
| Type | Trigger | Hold Duration | Behavior | Release Trigger |
|---|---|---|---|---|
| Sale Cart | Item added to POS cart | Until payment or void | Hard reserve. Other terminals see reduced available qty. | Payment completes (commit) or cart voided/abandoned (release). |
| Parked Transaction | Sale saved as parked | Until recalled or expired | Soft reserve. Other terminals see reduced available qty but with a visual warning: “2 units reserved by parked sale P-0045.” Staff can still sell through the soft reserve if they override the warning. | Parked transaction recalled and completed (commit), or voided (release), or expired after configurable timeout (release). |
| Transfer | Transfer approved and picking starts | Until transfer shipped or cancelled | Hard reserve at source location. Items being picked for an outbound transfer are reserved to prevent them from being sold before they ship. | Transfer shipped (status moves to IN_TRANSIT) or transfer cancelled (release). |
| Online Order | Online order placed and allocated to nearest store | Until fulfilled or cancelled | Hard reserve at the assigned store. The nearest-store allocation algorithm (see Module 1, Section 1.10 if applicable) assigns the order to the store with the most available stock. | Order fulfilled (commit) or order cancelled (release). |
| Hold-for-Pickup | Staff places a hold for a customer | Configurable expiry (default: 48 hours) | Hard reserve with auto-release on expiry. Customer has a window to pick up. If not picked up, system auto-releases the hold and notifies the store. | Customer picks up (commit), or expiry timer elapses (auto-release), or staff manually releases. |
Reservation Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
product_id | UUID | Yes | FK to product |
variant_id | UUID | No | FK to variant (if applicable) |
location_id | UUID | Yes | Location where the stock is reserved |
qty | Integer | Yes | Quantity reserved |
type | Enum | Yes | SALE_CART, PARKED_TRANSACTION, TRANSFER, ONLINE_ORDER, HOLD_FOR_PICKUP |
status | Enum | Yes | ACTIVE, COMMITTED, RELEASED, EXPIRED |
source_document_id | UUID | Yes | FK to the source document (sale ID, parked transaction ID, transfer ID, online order ID, or hold ID) |
source_document_type | String | Yes | Type of source document for polymorphic FK resolution |
reserved_by | UUID | Yes | User who created the reservation |
reserved_at | DateTime | Yes | Timestamp when the reservation was created |
expires_at | DateTime | No | Expiry timestamp for time-limited reservations (parked transactions, hold-for-pickup). Null for reservations without expiry. |
committed_at | DateTime | No | Timestamp when the reservation was committed (sale completed, transfer shipped). |
released_at | DateTime | No | Timestamp when the reservation was released (void, cancel, expiry). |
release_reason | String | No | Reason for release: VOID, CANCEL, EXPIRY, OVERRIDE |
tenant_id | UUID | Yes | Owning tenant |
Reservation Lifecycle State Machine
stateDiagram-v2
[*] --> ACTIVE: Reserve Created
ACTIVE --> COMMITTED: Sale Paid / Transfer Shipped / Pickup Completed
ACTIVE --> RELEASED: Void / Cancel / Manual Release
ACTIVE --> EXPIRED: Expiry Timer Elapsed
COMMITTED --> [*]
RELEASED --> [*]
EXPIRED --> [*]
note right of ACTIVE
Qty moved from AVAILABLE to RESERVED
Other terminals see reduced available qty
end note
note right of COMMITTED
Qty decremented from inventory
Reservation fulfilled
end note
note right of RELEASED
Qty moved from RESERVED back to AVAILABLE
Stock returned to sellable pool
end note
note right of EXPIRED
Auto-triggered by background job
Qty returned to AVAILABLE
Notification sent to staff
end note
Business Rules:
- When a reservation is created, the system atomically decrements the AVAILABLE status qty and increments the RESERVED status qty for the product-variant-location combination.
- When a reservation is committed, the RESERVED qty is decremented (stock leaves inventory via sale, or moves to IN_TRANSIT for transfer).
- When a reservation is released or expired, the RESERVED qty is decremented and the AVAILABLE qty is incremented (stock returns to sellable pool).
- Reservations are checked by a background job every 5 minutes for expiry. Expired reservations are auto-released and a notification is sent to the staff member who created the reservation.
- Parked transaction override: If a staff member at another terminal attempts to sell a product that has units reserved by a parked transaction, the system shows a warning: “2 of 5 units reserved by parked sale P-0045 at Terminal 2. Proceed anyway?” If the staff member confirms, the system sells through the available stock and the parked transaction’s reserved quantity is reduced when it is recalled (the system reconciles at recall time).
- Concurrent reservation conflict: If two terminals attempt to reserve the last available unit simultaneously, the first transaction to commit the database write wins. The second terminal receives an error: “Insufficient available stock. 0 units available.” This is enforced by database-level row locking on the inventory status record.
4.2.4 Minimum Display Quantity
Retail clothing stores rely on visual merchandising – an empty rack or sparse display reduces sales. The minimum display quantity feature provides an advisory warning system that alerts store staff when the available inventory at a location drops below a configured floor display minimum.
Key Behaviors:
- Minimum display quantity is configured per product (or variant) per location. Not all products require a minimum display – the field is optional and defaults to null (no warning).
- The warning is advisory only. It does not block sales, transfers, or any other operation. It is a soft alert that appears on the store dashboard and in the inventory list view.
- When available qty at a location drops below the configured minimum display qty, the system creates a dashboard alert: “Product XYZ at Store A has 1 unit remaining (minimum display: 3). Consider replenishment.”
- The alert clears automatically when stock is replenished above the minimum display qty (via receiving, transfer, or adjustment).
- Minimum display qty is distinct from the reorder point (Section 4.5). The reorder point triggers purchase order generation. The minimum display qty triggers a store-level visual merchandising alert.
Minimum Display Quantity Data Model
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | FK to product |
variant_id | UUID | No | FK to variant (if applicable). When set, the min display applies to the specific variant. When null, it applies to the product aggregate. |
location_id | UUID | Yes | FK to location. Min display is set per location since different stores may have different display requirements. |
min_display_qty | Integer | Yes | The minimum number of units that should be on display at this location. |
is_active | Boolean | Yes | Whether this minimum display rule is active. Allows temporary disabling without deleting the configuration. |
set_by | UUID | Yes | User who configured the minimum display qty. |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Minimum display quantity alerts appear on the store dashboard under a “Low Display Stock” section.
- Alerts are generated when
AVAILABLE qty < min_display_qtyat a location. The check runs whenever inventory changes at the location (sale, transfer, adjustment, receive). - Alerts include a suggested action: “Request transfer from [location with highest available qty]” with a one-click “Request Transfer” button.
- Minimum display quantity does NOT factor into the reorder point calculation (Section 4.5). They are independent systems.
- Setting a minimum display quantity of 0 is equivalent to disabling the alert for that product-location combination.
4.3 Purchase Orders & Procurement
Scope: Creating, approving, submitting, receiving, and closing purchase orders to replenish inventory from vendors. The PO workflow supports approval routing for high-value orders, partial receives, variance tracking, inspection steps, overdue alerts, and auto-generation from low-stock alerts.
4.3.1 Purchase Order State Machine
stateDiagram-v2
[*] --> DRAFT: PO Created
DRAFT --> PENDING_APPROVAL: Submit for Approval (above threshold)
DRAFT --> SUBMITTED: Submit to Vendor (below threshold / auto-approved)
PENDING_APPROVAL --> SUBMITTED: Manager Approves
PENDING_APPROVAL --> REJECTED: Manager Rejects
REJECTED --> DRAFT: Revise and Resubmit
SUBMITTED --> PARTIALLY_RECEIVED: Partial Shipment Arrived
PARTIALLY_RECEIVED --> PARTIALLY_RECEIVED: Additional Shipment
PARTIALLY_RECEIVED --> FULLY_RECEIVED: All Items Received
SUBMITTED --> FULLY_RECEIVED: Full Shipment Arrived
FULLY_RECEIVED --> CLOSED: PO Closed
DRAFT --> CANCELLED: Cancel Before Submit
SUBMITTED --> CANCELLED: Cancel After Submit
CANCELLED --> [*]
CLOSED --> [*]
note right of DRAFT
Editable line items
No inventory impact
Not sent to vendor
end note
note right of PENDING_APPROVAL
PO total exceeds approval threshold
Awaiting manager/owner approval
Line items locked for review
end note
note right of SUBMITTED
Sent to vendor (email/EDI/manual)
Awaiting shipment
Line items locked
end note
note right of PARTIALLY_RECEIVED
Some items received
Inventory incremented for received qty
Remaining items still expected
end note
note right of FULLY_RECEIVED
All line items received
Pending final review
Ready to close
end note
4.3.2 PO Approval Workflow
Purchase orders above a configurable dollar threshold require manager or owner approval before submission to the vendor. This prevents unauthorized large purchases while allowing routine restocking to flow without friction.
Approval Threshold Configuration:
| Setting | Type | Default | Description |
|---|---|---|---|
po_auto_approve_threshold | Decimal(10,2) | $2,000.00 | PO total value at or below this amount is auto-approved and moves directly to SUBMITTED. |
po_approval_role | Enum | MANAGER | Minimum role required to approve POs above threshold. Options: MANAGER, OWNER. |
po_approval_notify | Boolean | true | Whether to send push notification to approvers when a PO is pending. |
Approval Rules:
- When a staff member clicks “Submit” on a PO whose total value (
SUM of line_total) is at or below thepo_auto_approve_threshold, the PO moves directly from DRAFT to SUBMITTED. No approval step is needed. - When a staff member clicks “Submit” on a PO whose total value exceeds the
po_auto_approve_threshold, the PO moves from DRAFT to PENDING_APPROVAL. A notification is sent to all users with the configured approval role at the PO’s destination location. - The approver can APPROVE (moves to SUBMITTED), REJECT with a reason (moves to REJECTED), or request modifications (the PO creator is notified to revise).
- A REJECTED PO can be revised (returns to DRAFT status with editable line items) and resubmitted.
- The approval threshold is configurable per tenant in tenant settings. Different tenants may have different spending limits.
- Auto-generated draft POs from the reorder engine (Section 4.5) follow the same approval rules – they are not exempt from the threshold check.
Approval Workflow Sequence
sequenceDiagram
autonumber
participant S as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
participant NOTIF as Notification Service
participant M as Manager/Owner
S->>UI: Click "Submit PO"
UI->>API: POST /purchase-orders/{id}/submit
API->>DB: Calculate PO Total (SUM of line_total)
API->>DB: Lookup Tenant's po_auto_approve_threshold
alt PO Total <= Threshold (Auto-Approve)
API->>DB: Update Status: SUBMITTED
API->>DB: Lock Line Items
API-->>UI: "PO Submitted to Vendor"
Note right of API: PO proceeds to vendor submission
else PO Total > Threshold (Requires Approval)
API->>DB: Update Status: PENDING_APPROVAL
API->>DB: Lock Line Items for Review
API->>NOTIF: Send Approval Request to Manager(s)
API-->>UI: "PO Sent for Manager Approval"
NOTIF-->>M: "PO #PO-2026-00042 ($4,500) awaiting your approval"
alt Manager Approves
M->>UI: Review PO -> Click "Approve"
UI->>API: POST /purchase-orders/{id}/approve
API->>DB: Update Status: SUBMITTED
API->>DB: Record approved_by, approved_at
API->>NOTIF: Notify Creator: "PO Approved"
NOTIF-->>S: "Your PO #PO-2026-00042 was approved"
else Manager Rejects
M->>UI: Review PO -> Click "Reject"
M->>UI: Enter Rejection Reason
UI->>API: POST /purchase-orders/{id}/reject
API->>DB: Update Status: REJECTED
API->>DB: Record rejection_reason
API->>NOTIF: Notify Creator: "PO Rejected"
NOTIF-->>S: "Your PO #PO-2026-00042 was rejected: reason"
end
end
4.3.3 Purchase Order Lifecycle
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant DB as DB
participant V as Vendor
Note over U, V: Step 1: Create Purchase Order
U->>UI: Click "New Purchase Order"
UI->>UI: Select Vendor from List
UI->>API: GET /vendors/{id}/products
API-->>UI: Return Vendor's Product Catalog with Vendor Costs
loop Add Line Items
U->>UI: Select Product
UI->>UI: Auto-Fill Vendor SKU, Vendor Cost
U->>UI: Enter Quantity Ordered
U->>UI: Set Expected Delivery Date
UI->>UI: Calculate Line Total (qty x unit_cost)
end
UI->>UI: Display PO Summary (line count, total cost)
U->>UI: Add Notes (optional)
U->>UI: Click "Save Draft"
UI->>API: POST /purchase-orders
API->>DB: Create PO Record (Status: DRAFT)
Note right of DB: Auto-generated PO Number: PO-2026-00042
API-->>UI: PO #PO-2026-00042 Created
Note over U, V: Step 2: Submit to Vendor
U->>UI: Review PO -> Click "Submit"
UI->>API: POST /purchase-orders/{id}/submit
Note over API: Approval check runs here (see Section 4.3.2)
alt Email Submission
API->>V: Send PO via Email (PDF attachment)
Note right of V: Vendor receives PO email
else EDI Submission
API->>V: Transmit PO via EDI
else Manual Submission
API-->>UI: "PO marked Submitted - send manually"
Note right of UI: Staff prints PO and calls/faxes vendor
end
API->>DB: Update Status: SUBMITTED
API->>DB: Lock Line Items (no edits)
API-->>UI: PO Submitted Successfully
Note over U, V: Step 3: Receive Inventory
V-->>U: Shipment Arrives at Store/Warehouse
U->>UI: Open PO #PO-2026-00042 -> Click "Receive"
loop Receive Line Items
U->>UI: Enter Qty Received per Line Item
opt Variance Detected
UI-->>U: "Ordered: 50, Received: 48 - Enter Variance Note"
U->>UI: Enter Note: "2 units damaged in transit"
end
end
U->>UI: Click "Confirm Receive"
UI->>API: POST /purchase-orders/{id}/receive
par Inventory Updates
API->>DB: Increment Inventory (received qty per location)
API->>DB: Update PO Line Items (qty_received)
API->>DB: Record Variance Notes
end
alt All Items Received
API->>DB: Update Status: FULLY_RECEIVED
API-->>UI: "All items received"
else Partial Receive
API->>DB: Update Status: PARTIALLY_RECEIVED
API-->>UI: "Partial receive recorded - awaiting remaining"
end
Note over U, DB: Step 4 (Optional): Inspect Received Goods
opt Quality Inspection
U->>UI: Open Received Items -> Click "Inspect"
U->>UI: Mark Items as Passed / Failed
Note right of UI: Failed items logged for vendor claim
UI->>API: POST /purchase-orders/{id}/inspection
API->>DB: Record Inspection Results
end
Note over U, DB: Step 5: Close PO
U->>UI: Click "Close PO"
UI->>API: POST /purchase-orders/{id}/close
API->>DB: Update Status: CLOSED
API->>DB: Finalize Cost Records
API-->>UI: PO Closed
4.3.4 PO Header Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
po_number | String | Yes | Auto-generated: PO-{YEAR}-{SEQ} per tenant |
vendor_id | UUID | Yes | FK to vendor |
status | Enum | Yes | DRAFT, PENDING_APPROVAL, SUBMITTED, PARTIALLY_RECEIVED, FULLY_RECEIVED, CLOSED, CANCELLED, REJECTED |
destination_location_id | UUID | Yes | FK to the location where goods will be received |
total_value | Decimal(12,2) | Yes | Calculated: SUM of all line_total values |
expected_delivery_date | Date | No | Overall expected delivery date for the PO |
overdue_alert_buffer_days | Integer | No | Number of buffer days after expected_delivery_date before overdue alert triggers. Default: 3 days. |
submission_method | Enum | No | EMAIL, EDI, MANUAL. How the PO is sent to the vendor. |
notes | Text | No | Free-text notes for the PO |
auto_generated | Boolean | Yes | Whether this PO was auto-generated by the reorder engine (Section 4.5). Default: false. |
approved_by | UUID | No | Manager who approved the PO (if approval was required) |
approved_at | DateTime | No | Timestamp of approval |
rejection_reason | Text | No | Reason for rejection (if rejected) |
created_by | UUID | Yes | Staff member who created the PO |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
closed_at | DateTime | No | Timestamp when PO was closed |
4.3.5 PO Line Item Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Line item primary key |
purchase_order_id | UUID | Yes | Reference to parent PO |
product_id | UUID | Yes | Reference to product being ordered |
variant_id | UUID | No | Reference to specific variant (if applicable) |
vendor_sku | String | No | Vendor’s SKU (auto-filled from vendor-product link) |
qty_ordered | Integer | Yes | Quantity ordered from vendor |
qty_received | Integer | Yes | Quantity received so far (starts at 0) |
unit_cost | Decimal(10,2) | Yes | Cost per unit from vendor |
line_total | Decimal(10,2) | Yes | Calculated: qty_ordered x unit_cost |
expected_date | Date | No | Expected delivery date for this line |
received_date | Date | No | Actual date items were received |
variance_notes | String | No | Notes on quantity/quality discrepancies |
inspection_status | Enum | No | PENDING, PASSED, FAILED |
4.3.6 Expected Delivery Date & Overdue Alerts
Each purchase order has an expected_delivery_date field representing when the vendor is expected to deliver the goods. The system uses this date, combined with a configurable buffer period, to generate overdue alerts when a PO has not been received within the expected timeframe.
Overdue Alert Logic:
overdue_trigger_date = expected_delivery_date + overdue_alert_buffer_days
- If the current date exceeds
overdue_trigger_dateand the PO status is stillSUBMITTED(nothing received), the system generates an overdue alert. - If the PO status is
PARTIALLY_RECEIVEDand the current date exceedsoverdue_trigger_date, the system generates a different alert: “PO partially received but remaining items overdue.” - Overdue alerts appear on the purchasing dashboard and are sent as push notifications to the PO creator and the destination location’s manager.
- The
overdue_alert_buffer_daysdefaults to 3 days but can be overridden per PO when the expected delivery date is uncertain (e.g., international shipments). - Overdue alerts auto-clear when the PO reaches
FULLY_RECEIVEDorCLOSEDstatus.
Business Rules:
- If
expected_delivery_dateis not set on the PO, the system falls back to the vendor’s defaultlead_time_days(from the vendor record) plusoverdue_alert_buffer_days. - Overdue POs appear in the Open PO Report with a visual indicator (red highlight) and are sorted to the top by default.
- The background job that checks for overdue POs runs daily at a configurable time (default: 8:00 AM local time per tenant timezone).
4.3.7 Purchase Order Features
- Auto-generate PO from low-stock alerts: When inventory drops below
reorder_pointat any location, the system generates a suggested draft PO with the primary vendor and recommended quantities based on sales velocity (see Section 4.5 for reorder management details). - PO templates: Staff can save frequently ordered product sets as templates (e.g., “Weekly Nike Restock”) and generate new POs from templates with one click. Templates store vendor, product list, and default quantities but not dates or notes.
- Partial receives: Each receive operation records the quantity received per line item. Multiple receives accumulate until all items arrive. Each partial receive increments inventory at the destination location immediately.
- Variance tracking: When received quantity differs from ordered quantity, staff must enter a variance note. Variances are tracked for vendor performance reporting (see Section 4.3.8).
- Auto-increment PO number per tenant: PO numbers follow the format
PO-{YEAR}-{SEQUENCE}and auto-increment per tenant. Example:PO-2026-00001,PO-2026-00002. Sequence resets annually. - Receive to specific location: When receiving, staff selects the destination location (store or warehouse). Inventory increments at that location. The destination is pre-filled from the PO header’s
destination_location_idbut can be overridden during receiving. - PO duplication: Staff can duplicate an existing PO (any status) to create a new DRAFT with the same vendor and line items. Useful for recurring orders.
- Approval threshold: POs above a configurable dollar threshold require manager approval before vendor submission (see Section 4.3.2).
4.3.8 Reports: Purchase Orders
| Report | Purpose | Key Data Fields |
|---|---|---|
| Open PO Report | Track all non-closed purchase orders | PO number, vendor, status, total value, expected date, days since submitted, overdue flag |
| PO Receiving Report | Monitor receiving activity | PO number, line items, qty ordered vs received, variance %, receive date |
| Vendor Lead Time Report | Measure actual vs expected delivery | Vendor, PO count, avg expected lead time, avg actual lead time, on-time %, overdue count |
| PO Variance Report | Track quantity and quality discrepancies | PO number, line item, qty ordered, qty received, variance, variance notes |
| Cost Analysis Report | Review purchasing spend | Vendor, total PO value, product categories, avg unit cost, cost trends over time |
| Approval Pipeline Report | Monitor POs awaiting approval | PO number, total value, created by, created date, days pending, approver assigned |
| Overdue PO Report | Track POs past expected delivery | PO number, vendor, expected date, days overdue, last contact notes, status |
4.4 Receiving & Inspection
Scope: A single unified receiving workflow handles all inbound inventory regardless of source type – PO shipments, inter-store transfers, customer returns, vendor RMA replacements, and non-PO receives. This section documents the open receive mode, discrepancy handling, non-PO receiving, over-shipment handling, return-to-stock processing, and scanner-primary receiving operations.
4.4.1 Receiving Source Types
| Source Type | Origin | Example |
|---|---|---|
PO_RECEIVE | Purchase order from vendor | PO-2026-00042 shipment arrives |
TRANSFER_RECEIVE | Inter-store transfer | Transfer from Store A received at Store B |
RETURN_TO_STOCK | Customer return | Returned item added back to inventory |
RMA_REPLACEMENT | Vendor RMA replacement | Vendor sent replacement items |
NON_PO_RECEIVE | Stock received without a PO | Vendor sample, found stock, replacement |
4.4.2 Receiving Data Model – Header
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
receive_number | String | Yes | Auto-generated: RCV-{YEAR}-{SEQ} |
source_type | Enum | Yes | PO_RECEIVE, TRANSFER_RECEIVE, RETURN_TO_STOCK, RMA_REPLACEMENT, NON_PO_RECEIVE |
source_document_id | UUID | Conditional | FK to source document (PO, transfer, return, RMA). Required for all types except NON_PO_RECEIVE. |
non_po_reason_code | Enum | Conditional | Required when source_type = NON_PO_RECEIVE. See Section 4.4.6. |
location_id | UUID | Yes | Destination location where stock is received |
status | Enum | Yes | PENDING, IN_PROGRESS, COMPLETED |
received_by | UUID | Yes | Staff member processing the receive |
notes | Text | No | General notes for the receiving session |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
completed_at | DateTime | No | Timestamp when receiving was completed |
4.4.3 Receiving Data Model – Line Items
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
receive_id | UUID | Yes | Reference to parent receive record |
product_id | UUID | Yes | Reference to product |
variant_id | UUID | No | Reference to specific variant (if applicable) |
expected_qty | Integer | Yes | Quantity expected from source document. 0 for non-PO receive lines. |
received_qty | Integer | Yes | Actual quantity received |
variance | Integer | Computed | Calculated: received_qty - expected_qty |
condition | Enum | Yes | GOOD, DAMAGED, WRONG_ITEM |
condition_notes | Text | No | Notes describing the condition (especially for DAMAGED or WRONG_ITEM) |
initial_status | Enum | Yes | Inventory status assigned on receive: AVAILABLE (default for GOOD), DAMAGED, PENDING_INSPECTION |
serial_numbers[] | String[] | No | Serial numbers captured (if serial tracked) |
lot_number | String | No | Lot/batch number (if lot tracked) |
notes | Text | No | Notes on received items |
4.4.4 Open Receive Mode
Open receive mode is the primary workflow for receiving inventory against a purchase order. Staff sees the expected quantities from the PO and records the actual received quantities. Variances are automatically calculated and documented.
Workflow:
- Staff opens the PO in the receiving screen and clicks “Start Receiving.”
- The system displays all PO line items with their
qty_orderedand currentqty_received(from any prior partial receives). - For each line item, staff enters (or scans – see Section 4.4.9) the actual quantity received in this shipment.
- The system calculates the variance for each line:
received_qty_this_session - (qty_ordered - qty_previously_received). - If the variance is negative (short-shipped), the system highlights the line and prompts for a variance note.
- If the variance is positive (over-shipped), the system applies over-shipment rules (see Section 4.4.7).
- Staff confirms the receive. Inventory is incremented at the destination location for all GOOD items. DAMAGED items are placed in DAMAGED status. WRONG_ITEM items are flagged for RMA processing.
Open Receive Sequence Diagram
sequenceDiagram
autonumber
participant U as Staff
participant UI as Receiving UI
participant API as Backend
participant DB as DB
U->>UI: Open PO #PO-2026-00042
UI->>API: GET /purchase-orders/{id}/lines
API-->>UI: Return Line Items with Expected Qty
U->>UI: Click "Start Receiving"
UI->>API: POST /receiving/start
API->>DB: Create Receive Record (Status: IN_PROGRESS)
API-->>UI: Receive Session #RCV-2026-00108 Started
Note over U, UI: Staff sees expected qty for each line
loop Receive Each Line Item
alt Scanner Mode (Default)
U->>UI: Scan Item Barcode
UI->>UI: Match to PO Line Item
UI->>UI: Increment Received Qty by 1
Note right of UI: Each scan = +1 unit
else Manual Entry
U->>UI: Enter Received Qty for Line
end
UI->>UI: Calculate Variance (received - expected remaining)
opt Short-Shipped (Negative Variance)
UI-->>U: "Expected: 50, Received: 48 — 2 units short"
U->>UI: Enter Variance Note
end
opt Over-Shipped (Positive Variance)
UI-->>U: "Expected: 50, Received: 55 — 5 units over"
Note right of UI: Over-shipment rules apply (Section 4.4.7)
end
opt Damaged Item Found
U->>UI: Mark Item as DAMAGED
U->>UI: Enter Condition Notes
end
opt Wrong Item Found
U->>UI: Mark Item as WRONG_ITEM
U->>UI: Enter Condition Notes
end
end
U->>UI: Click "Confirm Receive"
UI->>API: POST /receiving/{id}/confirm
par Post-Receive Updates
API->>DB: Increment AVAILABLE Inventory (GOOD items)
API->>DB: Set DAMAGED items to DAMAGED Status
API->>DB: Flag WRONG_ITEM for RMA Processing
API->>DB: Update PO Line Items (qty_received)
API->>DB: Record Variance Notes
API->>DB: Log Movement Records (RECEIVE movement type)
API->>DB: Update Receive Status: COMPLETED
end
alt All PO Items Now Received
API->>DB: Update PO Status: FULLY_RECEIVED
API-->>UI: "PO fully received"
else Remaining Items Outstanding
API->>DB: Update PO Status: PARTIALLY_RECEIVED
API-->>UI: "Partial receive recorded — awaiting remaining items"
end
4.4.5 Discrepancy Handling (Triple Approach)
When receiving reveals discrepancies between expected and actual quantities or conditions, the system applies a triple approach to ensure nothing falls through the cracks:
- Note variance on PO line (always): Every variance is recorded on the PO line item’s
variance_notesfield with the quantity difference and the staff member’s explanation. This is mandatory for all discrepancies regardless of type. - Auto-create RMA draft for wrong/defective items: When items are marked as
WRONG_ITEMorDAMAGEDwith a condition indicating a vendor fault (not transit damage), the system auto-creates an RMA draft record linked to the PO and the vendor. The RMA draft appears in the returns management queue for staff to review and submit to the vendor. - Quarantine damaged items: Items marked as
DAMAGEDduring receiving are placed in theDAMAGEDinventory status at the receiving location. They are blocked from sale and transfer until resolved (write-off or vendor RMA).
Discrepancy Decision Flowchart
flowchart TD
A[Receive Line Item] --> B{Qty Matches Expected?}
B -->|Yes| C{Condition OK?}
B -->|No - Short| D[Note Variance on PO Line]
B -->|No - Over| E[Apply Over-Shipment Rules\nSection 4.4.7]
D --> F{All Items in Good Condition?}
F -->|Yes| G[Accept Short Shipment\nRecord Variance Note]
F -->|No| H{What Condition?}
C -->|Yes - GOOD| I[Accept to AVAILABLE Status]
C -->|No| H
H -->|DAMAGED| J[Move to DAMAGED Status]
H -->|WRONG_ITEM| K[Flag for RMA]
J --> L{Vendor Fault?}
L -->|Yes| M[Auto-Create RMA Draft\nLinked to PO & Vendor]
L -->|No - Transit Damage| N[Record Damage Note\nFile Carrier Claim if Applicable]
K --> M
E --> O{Within Over-Shipment Threshold?}
O -->|Yes - Accept| P[Accept Overage\nNote Variance]
O -->|No - Above Threshold| Q[Require Manager Approval\nto Accept Overage]
G --> R[Log Movement Record]
I --> R
M --> R
N --> R
P --> R
Q --> R
Business Rules:
- Every discrepancy, regardless of type, generates a variance note on the PO line item. This is non-negotiable.
- RMA drafts are auto-created only for vendor-attributable issues (wrong item, defective item). Transit damage is handled separately through carrier claims.
- Damaged items are immediately placed in DAMAGED status. They do not count toward the PO’s “received in good condition” tally.
- The PO Variance Report (Section 4.3.8) aggregates all discrepancies for vendor performance analysis.
- Discrepancy records include: PO number, line item, expected qty, received qty, variance, condition, notes, and whether an RMA was auto-created.
4.4.6 Non-PO Receiving
In some situations, stock arrives at a location without an associated purchase order. The system supports receiving without a PO, provided the staff member selects a mandatory reason code explaining why the stock is being received outside the normal procurement workflow.
Non-PO Reason Codes:
| Reason Code | Description | Example |
|---|---|---|
VENDOR_SAMPLE | Vendor sent sample merchandise for evaluation | New season sample box from Nike |
REPLACEMENT | Vendor sent replacement for previously defective/returned items outside the RMA process | Vendor shipped replacement directly without formal RMA |
RETURN_TO_STOCK | Items being re-entered into inventory after being temporarily removed (not a customer return – use RETURN_TO_STOCK source type for that) | Items returned from a photo shoot or trade show |
FOUND_STOCK | Stock discovered that was not in the system (e.g., found in back room, mislabeled) | Unscanned box found during warehouse cleanup |
OTHER | None of the above. Requires free-text explanation in notes field. | Unusual circumstance requiring documentation |
Non-PO Receive Data Model
Non-PO receives use the same receiving header and line item data models (Section 4.4.2 and 4.4.3) with the following specifics:
source_type=NON_PO_RECEIVEsource_document_id= nullnon_po_reason_codeis required (one of the codes above)expected_qtyon line items is set to 0 (since there is no source document to establish expectations)- Variance is not calculated for non-PO receives (there is no expected baseline)
Business Rules:
- Non-PO receives always require a reason code. The system rejects a non-PO receive without a reason code.
- When
reason_code = OTHER, the notes field on the receive header becomes mandatory. Staff must provide a free-text explanation. - Non-PO receives are flagged in the Receiving Log report for visibility. Management can filter the report to show only non-PO receives for audit purposes.
- Inventory incremented via non-PO receiving is costed at the product’s current weighted average cost (from the catalog). If no cost data exists, the system prompts staff to enter a unit cost.
4.4.7 Over-Shipment Handling
When a vendor ships more units than ordered, the system applies configurable rules to determine whether the overage is automatically accepted or requires manager approval.
Over-Shipment Configuration:
| Setting | Type | Default | Description |
|---|---|---|---|
over_shipment_threshold_pct | Decimal(5,2) | 10.00 | Maximum percentage above ordered qty that can be auto-accepted. |
over_shipment_approval_role | Enum | MANAGER | Role required to approve over-shipments above the threshold. |
Business Rules:
- Within threshold: If the received quantity exceeds the ordered quantity by up to the configured threshold percentage, the overage is auto-accepted. Inventory is incremented for the full received quantity. The variance is noted on the PO line item.
- Example: Ordered 100 units, threshold is 10%. Receiving up to 110 units is auto-accepted.
- Above threshold: If the received quantity exceeds the ordered quantity by more than the configured threshold percentage, the system blocks acceptance of the overage and requires manager approval.
- Example: Ordered 100 units, threshold is 10%. Receiving 115 units triggers manager approval for the 15-unit overage.
- Until approved, the over-threshold units are held in
PENDING_INSPECTIONstatus. The units within the threshold (110) are accepted immediately.
- Manager approval flow: The manager receives a notification: “Over-shipment on PO #PO-2026-00042, line 3: Ordered 100, Received 115 (15% over, threshold 10%). Approve acceptance?” The manager can approve (units move to AVAILABLE) or reject (units are flagged for return to vendor).
- Per-line calculation: The threshold is applied per line item, not per PO total. Each line item’s overage is evaluated independently.
- Cost impact: Over-shipped units accepted at the same unit cost as the PO line item. The PO total value is recalculated to reflect the actual received quantity.
4.4.8 Return-to-Stock
When a customer returns merchandise (processed through Module 1, Sales – Returns), the returned items re-enter the inventory system through the return-to-stock workflow.
Default Behavior:
- Customer returns automatically move to
AVAILABLEstatus. No inspection is required by default. - The rationale: clothing returns in this retail context are typically tried-on garments, not defective products. The default assumption is that returned items are saleable.
- Staff has the option to mark any returned item as
DAMAGEDduring the return process if the item is visibly damaged, soiled, or otherwise unsaleable.
Business Rules:
- Return-to-stock creates a receiving record with
source_type = RETURN_TO_STOCKandsource_document_idpointing to the return/refund transaction. - The returned item is added to inventory at the location where the return was processed (the store where the customer brought the item back).
- If the item is marked DAMAGED during return, it enters
DAMAGEDstatus instead ofAVAILABLE. The staff member must enter a condition note. - Return-to-stock inventory is costed at the original sale’s cost basis (from the sale transaction), not at current weighted average cost. This ensures accurate margin reporting.
- A
RETURN_TO_STOCKmovement record is logged to the audit trail.
4.4.9 Scanner-Primary Receiving
The default receiving workflow is scanner-primary: staff uses a barcode scanner to scan each individual item as it is unpacked. Each scan auto-increments the received count for the matching PO line item by one unit.
Workflow:
- Staff opens the PO receiving screen and clicks “Start Receiving.”
- The system enters scanner mode (default). The cursor focus is on the barcode scan input field.
- Staff scans an item’s barcode. The system:
- Looks up the barcode in the catalog (Module 3).
- Matches it to a PO line item.
- Increments the
received_qtyfor that line by 1. - Plays an audible confirmation beep.
- Displays a running count: “Item XYZ: 23 of 50 received.”
- If the barcode does not match any PO line item, the system shows an alert: “Barcode not found on this PO. Wrong item?” Staff can flag it as WRONG_ITEM or search manually.
- Staff repeats scanning until all items are processed.
- Staff clicks “Confirm Receive” to finalize.
Manual Override:
- For items with damaged or missing barcodes, staff can switch to manual entry mode for individual line items.
- In manual mode, staff selects the product from the PO line item list and enters the quantity directly.
- The system logs whether each line was received via scanner or manual entry (for accuracy tracking).
Business Rules:
- Scanner mode is the default. The receiving screen opens in scanner mode unless the staff member explicitly switches to manual.
- Each barcode scan increments the count by exactly 1. There is no “scan and enter quantity” mode in scanner-primary workflow – every physical unit is scanned individually.
- If the same barcode is scanned more than the expected quantity for that line, over-shipment rules (Section 4.4.7) apply.
- Scanning speed is optimized for high throughput: the system processes each scan in under 200ms and immediately updates the on-screen count.
- The receiving screen shows a progress summary at all times: total items expected, total scanned so far, and lines remaining.
4.4.10 Unified Receiving Sequence (All Source Types)
sequenceDiagram
autonumber
participant U as Staff
participant UI as Receiving UI
participant API as Backend
participant DB as DB
Note over U, DB: Unified Receiving Flow
U->>UI: Select Source Document (PO / Transfer / Return / RMA / Non-PO)
UI->>API: GET /receiving/source/{type}/{id}
API-->>UI: Return Expected Line Items (empty for Non-PO)
U->>UI: Click "Start Receiving"
API->>DB: Create Receive Record (Status: IN_PROGRESS)
loop Receive Each Line Item
alt Scanner Verification (Default)
U->>UI: Scan Item Barcode
UI->>UI: Match to Expected Line Item
UI->>UI: Increment Received Qty
else Manual Entry
U->>UI: Enter Received Qty per Line
end
opt Variance Detected
UI-->>U: "Expected: 50, Received: 48"
U->>UI: Enter Variance Notes
end
opt Damaged Item
U->>UI: Mark Condition: DAMAGED
U->>UI: Enter Damage Notes
end
opt Wrong Item
U->>UI: Mark Condition: WRONG_ITEM
U->>UI: Enter Notes
end
opt Serial Tracked Product
U->>UI: Scan/Enter Serial Number for Each Unit
end
opt Lot Tracked Product
U->>UI: Enter Lot Number
end
end
U->>UI: Click "Confirm Receive"
UI->>API: POST /receiving/{id}/confirm
par Post-Receive Updates
API->>DB: Increment Inventory at Location (Received Qty - GOOD → AVAILABLE)
API->>DB: Set Damaged Items to DAMAGED Status
API->>DB: Flag Wrong Items for RMA Processing
API->>DB: Update Source Document (PO/Transfer/Return/RMA status)
API->>DB: Log Movement Records (per source type)
API->>DB: Update Receive Status: COMPLETED
end
API-->>UI: Receiving Complete
4.4.11 Reports: Receiving & Inspection
| Report | Purpose | Key Data Fields |
|---|---|---|
| Receiving Log | All receiving activity across all source types | Receive number, source type, source document, location, items expected, items received, variances, receive date |
| Non-PO Receiving Report | Track inventory received outside of PO process | Receive number, reason code, products, qty, staff member, date, notes |
| Damaged Goods Report | Track items received in damaged condition | Receive number, source, product, qty damaged, condition notes, RMA created (Y/N) |
| Over-Shipment Report | Track vendor over-shipments and approval outcomes | PO number, line item, qty ordered, qty received, overage %, auto-accepted (Y/N), approval status |
| Receiving Accuracy Report | Measure scanner vs manual receive accuracy | Location, total items received, scanner-received count, manual-received count, variance rate by method |
4.5 Reorder Management
Scope: Automating inventory replenishment through velocity-based reorder points, seasonal demand adjustments, and auto-generated draft purchase orders for staff review and approval. The reorder engine reduces stockouts by proactively identifying when products need replenishment and pre-building purchase orders for staff to review. This section also covers static override of reorder points and dead stock detection.
4.5.1 Velocity-Based Reorder Points
The system calculates dynamic reorder points per product per location using sales velocity, vendor lead time, and configurable safety stock. This ensures that reorder triggers adapt automatically to changing demand patterns without requiring manual intervention.
Formula:
reorder_point = (avg_daily_sales x lead_time_days x seasonal_factor) + safety_stock
Components:
| Component | Source | Description |
|---|---|---|
avg_daily_sales | Calculated from rolling 90-day sales velocity per product per location. | The average number of units sold per day over the trailing 90 days. New products with less than 90 days of history use the available history. Products with zero sales use 0. |
lead_time_days | Sourced from vendor-product relationship (vendor default or per-product override). | The number of days between placing a PO with the vendor and receiving the goods. |
safety_stock | Configurable multiplier applied to the standard deviation of daily sales. Default: 1.5 sigma. | Buffer stock to account for demand variability and supply uncertainty. Higher sigma values provide more protection against stockouts at the cost of higher inventory. |
seasonal_factor | Multiplier derived from historical same-period data (e.g., December velocity in prior years). | Applied to avg_daily_sales before reorder point calculation. A factor of 1.5 means the system expects 50% higher sales than the rolling average suggests. Default: 1.00 (no seasonal adjustment). |
Recalculation: Weekly via background job (configurable schedule per tenant). The job recalculates reorder points for all active products at all locations and updates the reorder point data model.
Reorder Point Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
product_id | UUID | Yes | Reference to product |
location_id | UUID | Yes | Reference to store/warehouse location |
avg_daily_velocity | Decimal(8,3) | Yes | Rolling 90-day average daily sales |
lead_time_days | Integer | Yes | Vendor lead time for this product |
safety_stock_units | Integer | Yes | Calculated safety stock buffer |
reorder_point | Integer | Yes | Stock level that triggers reorder |
reorder_qty | Integer | Yes | Economic order quantity (recommended purchase qty) |
seasonal_factor | Decimal(5,2) | No | Seasonal adjustment multiplier (default: 1.00) |
override_reorder_point | Integer | No | Manager-set manual override. When not null, takes precedence over the calculated reorder_point. See Section 4.5.3. |
override_reason | Text | No | Documentation for why the override was set. Required when override_reorder_point is not null. |
override_set_by | UUID | No | User who set the override. |
override_set_at | DateTime | No | Timestamp when the override was set. |
last_calculated_at | DateTime | Yes | Timestamp of last recalculation |
tenant_id | UUID | Yes | Owning tenant |
4.5.2 Auto-Generated Draft POs
When stock at any location drops below the calculated reorder_point (or the override_reorder_point if set), the system creates a draft purchase order for staff review.
Auto-PO Features:
- Pre-filled with the primary vendor for each product below reorder point.
- Recommended quantity set to
reorder_qty(economic order quantity). - Vendor consolidation: If multiple products from the same vendor hit reorder simultaneously, the system combines them into a single draft PO. This reduces the number of POs and may help reach vendor minimum order values.
- Staff receives notification: “3 draft POs auto-generated for review.”
- Staff can edit quantities, add/remove products, then submit – or discard the draft.
- Auto-generated draft POs are marked with
auto_generated = trueon the PO header (Section 4.3.4) so they can be filtered and reported on separately. - Auto-generated POs follow the same approval threshold rules as manually created POs (Section 4.3.2). They are not exempt from the approval workflow.
Auto-PO Generation Sequence
sequenceDiagram
autonumber
participant JOB as Background Job
participant DB as DB
participant API as Backend
participant NOTIF as Notification Service
participant U as Staff
Note over JOB, U: Reorder Point Check (Runs on Schedule)
JOB->>DB: Query Products Below Reorder Point (or Override)
DB-->>JOB: Return List (Product, Location, Current Qty, Reorder Point)
JOB->>JOB: Exclude products with existing DRAFT/SUBMITTED PO for same vendor
loop For Each Product Below Reorder
JOB->>DB: Lookup Primary Vendor
JOB->>DB: Get Reorder Qty (Economic Order Quantity)
alt Existing Draft PO for Same Vendor
JOB->>DB: Add Line Item to Existing Draft PO
else No Existing Draft PO
JOB->>DB: Create New Draft PO for Vendor
JOB->>DB: Add Line Item
end
end
JOB->>DB: Finalize Draft POs (calculate totals)
JOB->>NOTIF: "3 draft POs auto-generated for review"
NOTIF-->>U: Push Notification / Dashboard Alert
Note over U, DB: Staff Review
U->>API: GET /purchase-orders?status=DRAFT&auto_generated=true
API-->>U: Return Auto-Generated Draft POs
alt Approve as-is
U->>API: POST /purchase-orders/{id}/submit
Note right of API: Approval threshold check applies (Section 4.3.2)
API->>DB: Update Status: SUBMITTED (or PENDING_APPROVAL)
else Modify and Submit
U->>API: PATCH /purchase-orders/{id}
Note right of U: Edit quantities, add/remove lines
U->>API: POST /purchase-orders/{id}/submit
else Discard
U->>API: DELETE /purchase-orders/{id}
API->>DB: Delete Draft PO
end
Business Rules:
- The reorder check job does not create duplicate draft POs. If a product already has an open draft or submitted PO for the same vendor, it is excluded from auto-PO generation.
- The reorder check evaluates the
override_reorder_pointfirst. If set, it uses the override value. If null, it uses the calculatedreorder_point. - Auto-PO generation uses the product’s AVAILABLE quantity (not total on-hand) to determine if reorder is needed. RESERVED, IN_TRANSIT, and other non-available statuses are excluded.
- The system accounts for in-transit stock when calculating whether to reorder. If a product has an open PO with expected delivery within the lead time window, the system may skip reorder to avoid over-ordering (configurable behavior:
account_for_open_possetting, default: true).
4.5.3 Static Override
In some cases, the velocity-based reorder point calculation does not reflect the manager’s knowledge of the business. For example, a product may have low sales velocity (suggesting a low reorder point) but the manager knows it will be featured in an upcoming promotion and needs extra stock. Or a product may have high velocity but the manager knows it is being discontinued and does not want to reorder.
The static override feature allows a manager to lock any product at any location to a manual reorder point that overrides the dynamic velocity calculation.
Behavior:
- When
override_reorder_pointis set (not null), the reorder engine uses this value instead of the calculatedreorder_point. - The calculated
reorder_pointcontinues to be recalculated weekly by the background job, but it is ignored for reorder triggering as long as the override is active. - The override is visible in the product’s inventory detail screen with a visual indicator: “Reorder point manually set to 25 by [Manager Name] on [Date]. Calculated value: 12.”
- Managers can remove the override at any time, returning the product to dynamic reorder point calculation.
Business Rules:
- Only users with MANAGER or OWNER role can set or remove reorder point overrides.
- When setting an override, the
override_reasonfield is mandatory. The manager must document why the override is being set (e.g., “Upcoming Black Friday promotion – need extra stock”, “Discontinuing product – do not reorder”). - Overrides are per product per location. A manager can override the reorder point at Store A without affecting the calculation at Store B.
- The Reorder Alerts report (Section 4.5.5) shows which products are using overridden reorder points vs. calculated reorder points, so management can audit active overrides.
- There is no expiry on overrides. They remain active until manually removed. A periodic review reminder can be configured (e.g., “3 reorder overrides have been active for 90+ days – review?”).
4.5.4 Dead Stock Detection
Dead stock (also called slow-moving or stagnant inventory) represents products that have not sold at a location for an extended period. These items tie up capital, occupy shelf space, and may need to be marked down, transferred to a higher-traffic store, or written off.
Detection Logic:
- The system monitors the sales velocity of every active product at every location.
- When a product’s sales velocity at a location is zero for a configurable number of consecutive days (default: 90 days), it is flagged as dead stock.
- The dead stock flag is recalculated by the same weekly background job that recalculates reorder points.
Alert Behavior:
- Dead stock items appear on the Dead Stock Report (dashboard and exportable).
- A dashboard alert is displayed: “15 products at Store A have had zero sales for 90+ days.”
- The alert is informational only. No automatic action is taken. The manager decides the appropriate action for each item.
Manager Actions (Manual):
| Action | Description |
|---|---|
| Markdown | Reduce the price to accelerate sales. Handled via Module 3 (Catalog) price update. |
| Transfer | Move stock to a location with higher demand. Handled via Module 5 (Transfers). |
| Write-Off | Remove from inventory as a loss. Handled via Section 4.7 (Adjustments) with reason code WRITE_OFF. |
| Dismiss Alert | Acknowledge the alert without taking action. The product remains flagged but the alert is silenced for a configurable period (default: 30 days). |
Business Rules:
- Dead stock detection only applies to products with
lifecycle_status = ACTIVEin the catalog. Discontinued or inactive products are excluded. - The configurable threshold (default 90 days) is set per tenant in tenant settings. Different tenants may have different thresholds based on their product turnover expectations.
- Dead stock is evaluated per location. A product may be dead stock at Store A but selling well at Store B. The system highlights this imbalance and suggests transfer as an action.
- Products that have been in inventory for less than the threshold period (e.g., a new product received 30 days ago) are excluded from dead stock detection regardless of sales velocity.
- The Dead Stock Report includes: product, location, current qty, last sale date, days since last sale, total value at cost, suggested action.
Dead Stock Data Model
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | FK to product |
location_id | UUID | Yes | FK to location |
days_since_last_sale | Integer | Yes | Number of days since the last sale of this product at this location |
last_sale_date | Date | No | Date of the most recent sale. Null if never sold at this location. |
qty_on_hand | Integer | Yes | Current AVAILABLE quantity at this location |
value_at_cost | Decimal(10,2) | Yes | qty_on_hand x weighted_avg_cost |
is_flagged | Boolean | Yes | Whether this product-location is currently flagged as dead stock |
flagged_at | DateTime | No | Timestamp when the dead stock flag was set |
alert_dismissed_at | DateTime | No | Timestamp when a manager dismissed the alert. Alert is silenced until dismissed_at + dismiss_duration. |
dismiss_duration_days | Integer | No | Number of days to silence the alert after dismissal. Default: 30. |
tenant_id | UUID | Yes | Owning tenant |
4.5.5 Reports: Reorder Management
| Report | Purpose | Key Data Fields |
|---|---|---|
| Reorder Alerts | Products below reorder point needing attention | Product, location, current qty, reorder point (calculated vs override), days of supply remaining, suggested qty, primary vendor, override active (Y/N) |
| Auto-PO Performance | Effectiveness of automatic reorder system | Auto-generated PO count, submitted as-is count, modified count, cancelled count, avg fill rate, avg time from draft to submit |
| Velocity Trends | Sales velocity changes over time per product | Product, 30-day velocity, 60-day velocity, 90-day velocity, trend direction (increasing/stable/decreasing), seasonal factor |
| Days of Supply | How long current stock will last at current velocity | Product, location, qty on hand, avg daily velocity, days of supply, reorder urgency (Critical/Low/OK) |
| Dead Stock Report | Products with zero velocity over threshold period | Product, location, current qty, last sale date, days since last sale, value at cost, suggested action, alert status |
| Override Audit Report | Active reorder point overrides for review | Product, location, override value, calculated value, override reason, set by, set date, days active |
4.6 Inventory Counting & Auditing
Scope: Maintaining inventory accuracy through structured counting workflows that reconcile system quantities with physical reality. The system supports five counting methods, two count modes (freeze and snapshot), scanner-primary counting, and a complete review-and-approve workflow for count variances.
4.6.1 Count Types
The system supports five counting methods to maintain inventory accuracy. Each method is suited to different operational needs.
| Method | Description | Frequency | Scope |
|---|---|---|---|
| Full Physical Count | All products at a location counted | Annual or semi-annual | Entire location |
| Cycle Count | Rolling partial counts by category | Weekly (configurable schedule) | Category rotation |
| Scanner-Assisted Count | Barcode/RFID scanner used to tally items | As needed | Configurable scope |
| Monthly Scan | Scheduled full-location scan | Auto-created 1st of each month (configurable) | Entire location |
| On-Demand Count | Ad-hoc count triggered by manager | As needed | Specific products |
4.6.2 Count Data Model – Header
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
count_number | String | Yes | Auto-generated: CNT-{YEAR}-{SEQ} |
type | Enum | Yes | FULL, CYCLE, SCANNER, MONTHLY, ON_DEMAND |
location_id | UUID | Yes | Location being counted |
status | Enum | Yes | CREATED, IN_PROGRESS, REVIEW, APPROVED, CANCELLED |
count_mode | Enum | Yes | FREEZE or SNAPSHOT. See Section 4.6.4. |
scope | Enum | Yes | ALL, CATEGORY, PRODUCT_LIST |
category_ids[] | UUID[] | No | Categories included (when scope = CATEGORY) |
product_ids[] | UUID[] | No | Specific products included (when scope = PRODUCT_LIST) |
snapshot_taken_at | DateTime | No | Timestamp when the inventory snapshot was taken (SNAPSHOT mode only). |
created_by | UUID | Yes | Manager who initiated the count |
assigned_to | UUID | No | Staff member assigned to perform the count |
approved_by | UUID | No | Manager who approved adjustments |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
completed_at | DateTime | No | Timestamp when count was approved/cancelled |
4.6.3 Count Data Model – Line Items
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
count_id | UUID | Yes | Reference to parent count |
product_id | UUID | Yes | Product being counted |
variant_id | UUID | No | Specific variant (if applicable) |
expected_qty | Integer | Yes | System’s recorded AVAILABLE quantity at count start (or snapshot time in SNAPSHOT mode) |
counted_qty | Integer | No | Actual quantity counted by staff |
variance | Integer | Computed | Calculated: counted_qty - expected_qty |
variance_pct | Decimal(5,2) | Computed | Calculated: (variance / expected_qty) x 100. Null if expected_qty is 0. |
count_method | Enum | No | SCANNER or MANUAL – indicates how this line was counted |
adjustment_approved | Boolean | No | Whether the variance adjustment was approved by the reviewing manager |
notes | Text | No | Staff notes on variance explanation |
4.6.4 Configurable Count Freeze
When initiating a stock count, the manager chooses one of two counting modes. Each mode has different trade-offs between accuracy and operational impact.
| Aspect | FREEZE Mode | SNAPSHOT Mode |
|---|---|---|
| POS Sales | Blocked at the counting location for the duration of the count. Other locations continue to sell normally. | Continue normally. Sales are recorded and reconciled after the count. |
| Transfers | Outbound transfers blocked. Inbound transfers held until count completes. | Continue normally. Transfers are recorded and reconciled after the count. |
| Accuracy | Highest accuracy – no inventory movement during count ensures perfect reconciliation. | Slightly lower accuracy – reconciliation must account for sales and transfers that occurred during the count window. |
| Business Impact | High – store cannot sell during count. Best for after-hours or slow periods. | Low – store operates normally. Suitable for counts during business hours. |
| Duration Limit | No system limit, but operational pressure to complete quickly since sales are blocked. | No limit. Count can span hours or even days if needed. |
| Reconciliation | Direct comparison: counted_qty vs expected_qty. No adjustment needed for concurrent activity. | System calculates: adjusted_expected = snapshot_qty - sales_during_count + receives_during_count. Variance = counted_qty - adjusted_expected. |
| Recommended For | Full physical counts, high-value inventory sections, annual audits. | Cycle counts, monthly scans, routine counting during business hours. |
Business Rules:
- The count mode is selected at count creation time and cannot be changed after the count starts (status = IN_PROGRESS).
- FREEZE mode: When a count is started in FREEZE mode, the POS at the counting location displays a message: “Inventory count in progress. Sales temporarily suspended at this location.” POS terminals at other locations are unaffected. Transfers to/from the counting location are queued and processed after the count is approved.
- SNAPSHOT mode: When a count is started in SNAPSHOT mode, the system takes a point-in-time snapshot of all in-scope product quantities. This snapshot is stored as the
expected_qtyfor each count line item. Sales and transfers that occur after the snapshot continue normally. At review time, the system recalculates the expected values by applying all inventory movements that occurred between the snapshot time and the count submission time. - Only users with MANAGER or OWNER role can create counts and select the count mode.
- The system defaults to SNAPSHOT mode. FREEZE mode must be explicitly selected.
4.6.5 Scanner-Primary Counting
The default counting workflow is scanner-primary: staff uses a barcode scanner to scan each physical item on the shelf. Each scan increments the count for the matching product by one unit.
Scanner-Assisted Count Workflow
sequenceDiagram
autonumber
participant M as Manager
participant U as Staff
participant SC as Scanner Device
participant UI as Count UI
participant API as Backend
participant DB as DB
Note over M, DB: Step 1: Create Count Session
M->>UI: Click "New Stock Count"
M->>UI: Select Type, Location, Scope, Mode
UI->>API: POST /stock-counts
API->>DB: Create Count Header (Status: CREATED)
alt SNAPSHOT Mode
API->>DB: Snapshot Current Qty for All In-Scope Products
Note right of DB: Snapshot stored as expected_qty per line
else FREEZE Mode
API->>DB: Set Location POS to Count Mode (sales blocked)
end
API->>DB: Create Count Line Items (one per in-scope product-variant)
API-->>UI: Count #CNT-2026-00031 Created
Note over M, DB: Step 2: Assign & Start Counting
M->>UI: Assign Count to Staff Member
API-->>U: Notification: "Count assigned to you"
U->>UI: Open Count -> Click "Start Counting"
API->>DB: Update Status: IN_PROGRESS
Note over U, SC: Step 3: Scan Items
loop For Each Physical Item on Shelf
U->>SC: Scan Item Barcode
SC-->>UI: Barcode Data
UI->>UI: Lookup Product by Barcode
UI->>UI: Increment counted_qty by 1 for Matching Line
alt Barcode Matched
UI-->>U: Beep + "Item XYZ: 23 counted"
else Barcode Not Found
UI-->>U: Alert "Unknown barcode — enter product manually?"
U->>UI: Search Product by SKU/Name
U->>UI: Confirm Match
UI->>UI: Increment counted_qty by 1
Note right of UI: Line marked as count_method = MANUAL
end
end
opt Items with Damaged/Missing Barcodes
U->>UI: Switch to Manual Entry for Specific Line
U->>UI: Enter counted_qty Directly
Note right of UI: Line marked as count_method = MANUAL
end
Note over U, DB: Step 4: Submit for Review
U->>UI: Click "Submit for Review"
UI->>API: POST /stock-counts/{id}/submit
API->>DB: Calculate Variances per Line
API->>DB: Update Status: REVIEW
Note over M, DB: Step 5: Review & Approve
M->>UI: Open Count for Review
UI->>API: GET /stock-counts/{id}/variances
API-->>UI: Return Variance Report
M->>UI: Review Each Variance Line
Note right of M: Accept or reject each line adjustment
M->>UI: Click "Approve Adjustments"
UI->>API: POST /stock-counts/{id}/approve
par Inventory Updates
API->>DB: Apply Approved Adjustments to Inventory
API->>DB: Log Each Adjustment as COUNT_ADJUST Movement
API->>DB: Update Count Status: APPROVED
end
alt FREEZE Mode
API->>DB: Release Location POS from Count Mode (sales resume)
API->>DB: Process Queued Transfers
end
API-->>UI: Count Approved — Inventory Updated
Business Rules:
- Scanner mode is the default. The count screen opens in scanner-listening mode when the count is started.
- Each barcode scan increments the count by exactly 1. Staff scans every physical unit individually.
- Manual quantity entry is available as a fallback for items with damaged or missing barcodes. Staff can switch between scanner and manual mode on a per-line basis.
- Items not scanned during the count have
counted_qty = 0at submission time. If the expected qty was greater than 0, this is flagged as a variance (potential shrinkage or miscount). - Products scanned that are not in the count scope (e.g., wrong category during a cycle count) are flagged with a warning but can be added to the count at the staff member’s discretion.
- Count line items track whether they were counted via scanner or manual entry (
count_methodfield) for accuracy auditing.
4.6.6 Count Workflow Sequence
sequenceDiagram
autonumber
participant M as Manager
participant U as Staff
participant UI as Inventory UI
participant API as Backend
participant DB as DB
Note over M, DB: Step 1: Create Count
M->>UI: Click "New Stock Count"
M->>UI: Select Type (Full/Cycle/On-Demand)
M->>UI: Select Location
M->>UI: Select Count Mode (Freeze/Snapshot)
M->>UI: Define Scope (All / Category / Product List)
UI->>API: POST /stock-counts
API->>DB: Create Count (Status: CREATED)
API->>DB: Snapshot Expected Qty for All In-Scope Products
API-->>UI: Count #CNT-2026-00031 Created
Note over M, DB: Step 2: Assign & Count
M->>UI: Assign Count to Staff Member
API-->>U: Notification: "Count assigned to you"
U->>UI: Open Count -> Start Counting
API->>DB: Update Status: IN_PROGRESS
loop Count Each Product
alt Scanner-Assisted (Default)
U->>UI: Scan Item Barcode
UI->>UI: Increment Counted Qty for Product
else Manual Entry (Fallback)
U->>UI: Enter Counted Qty per Product
end
end
U->>UI: Click "Submit for Review"
UI->>API: POST /stock-counts/{id}/submit
API->>DB: Update Status: REVIEW
Note over M, DB: Step 3: Review & Approve
M->>UI: Open Count for Review
UI->>API: GET /stock-counts/{id}/variances
API-->>UI: Return Variance Report
M->>UI: Review Each Variance
Note right of M: Accept or reject each line adjustment
M->>UI: Click "Approve Adjustments"
UI->>API: POST /stock-counts/{id}/approve
par Inventory Updates
API->>DB: Apply Approved Adjustments to Inventory
API->>DB: Log Each Adjustment as COUNT_ADJUST Movement
API->>DB: Update Status: APPROVED
end
API-->>UI: Count Approved - Inventory Updated
4.6.7 Reports: Inventory Counting
| Report | Purpose | Key Data Fields |
|---|---|---|
| Count Variance Report | Variances discovered during stock counts | Count number, type, mode (Freeze/Snapshot), product, expected qty, counted qty, variance, variance %, count method (Scanner/Manual), adjustment status |
| Count Schedule Report | Upcoming and overdue scheduled counts | Count type, location, scheduled date, status, assigned to, overdue flag |
| Count Accuracy Trend | Track counting accuracy over time | Month, location, total counts, avg variance %, scanner-counted %, manual-counted %, accuracy trend |
| Shrinkage by Location | Inventory loss detected through counts per location | Location, period, total negative variances, value at cost, shrinkage % of total inventory value |
4.6.8 RFID-Assisted Counting (Raptag)
RFID counting operates as a dedicated subsystem separate from barcode-scanner counting. While scanner-primary counting (Section 4.6.5) handles individual barcode scans at the POS, RFID counting enables bulk tag reads using handheld RFID readers via the Raptag mobile application (Chapter 16).
Scope: RFID counting is for inventory counts ONLY. It does not participate in receiving (Section 4.4), transfers (Section 4.5), or sales (Module 1). Configuration for the RFID subsystem is in Section 5.16.
RFID Count Session Lifecycle
Manager Creates Session → Assigns Sections to Operators → Operators Join via Raptag
→ Parallel Scanning (offline-capable) → Upload Chunks → Server Merges & Deduplicates
→ Variance Calculation → Manager Reviews → Approve / Recount
Multi-Operator Sessions
A single RFID count session can have multiple operators (up to 10), each assigned to a section of the store. This enables parallel counting for enterprise-scale inventories (100,000+ items).
Workflow:
- Manager creates a count session in the Admin Portal or Raptag app, selecting count type (full_inventory, cycle_count, spot_check) and location
- Manager assigns sections to operators (e.g., “Sarah: Men’s Tops”, “James: Women’s Bottoms”)
- Each operator launches Raptag, sees the active session on their Home Dashboard, and taps “Join Session”
- Each operator’s device scans independently using the Zebra reader, recording tag reads locally in SQLite (offline-capable)
- On upload, the system merges all operator reads into one session via chunked upload (5,000 events per chunk)
- Server-side deduplication: If two operators scan the same tag (same EPC), the system keeps the read with the strongest RSSI (closest proximity = most accurate location)
Multi-Operator Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to shared.tenants |
session_id | UUID | Yes | FK to rfid_scan_sessions |
operator_id | UUID | Yes | FK to shared.users |
assigned_section | Text | No | Section/area label assigned by manager |
device_id | UUID | No | FK to devices — reader device used by operator |
joined_at | Timestamp | Yes | When operator joined the session |
left_at | Timestamp | No | When operator left or session ended |
Deduplication Rules:
- Same EPC scanned by multiple operators → keep the read with highest RSSI (strongest signal indicates closest proximity)
- Merge happens server-side during chunk upload processing
rfid_scan_eventsrecords which operator reported each tag (viasession_operatorslink)
RFID-Specific Business Rules:
- Maximum 10 operators per session
- Each operator can only participate in one active session at a time
- Manager must create the session and assign at least one section before operators can join
- Session cannot be completed until all operators have submitted their chunks (or been removed)
- Operators can leave a session (voluntarily or by manager removal) without losing already-uploaded data
- If an operator’s device loses power mid-scan, auto-save preserves data locally; they can resume on the same or different device
RFID Counting vs Scanner Counting
| Aspect | Scanner-Primary (Section 4.6.5) | RFID-Assisted (This Section) |
|---|---|---|
| Input Device | Barcode scanner (USB HID) | Zebra RFID reader (MC3390R, RFD40) |
| Application | POS terminal | Raptag mobile app (.NET MAUI) |
| Read Speed | 1 item per scan | 40+ tags per second |
| Tracking Level | SKU-level (barcode → product lookup) | EPC-level (individual item tracking) |
| Offline | Connected to POS (online) | Fully offline with SQLite + sync |
| Operators | Single operator per count | Up to 10 operators per session |
| Data Flow | Real-time via POS API | Batch upload via chunked sync |
| Count Method Value | SCANNER or MANUAL | RFID |
4.7 Inventory Adjustments
Scope: Manual inventory adjustments handle corrections outside of stock counts. Adjustments are used when a staff member discovers a discrepancy between the system quantity and physical reality and needs to correct the system immediately, without waiting for a scheduled count. All adjustments – regardless of direction or size – require manager approval before stock levels change.
4.7.1 Adjustment Approval Workflow
All adjustments require manager approval. Unlike the previous threshold-based approach, this module enforces a universal approval requirement for every inventory adjustment. This ensures that no stock level change bypasses management oversight, which is critical for loss prevention and financial accuracy in a multi-store retail environment.
Workflow:
- A staff member identifies a discrepancy (e.g., found 3 extra units on a shelf, or 2 units are missing and believed stolen).
- The staff member creates an adjustment request in the system, specifying the product, location, quantity change (positive or negative), reason code, and notes.
- The adjustment is created with
approval_status = PENDING. Inventory is NOT changed yet. - A notification is sent to all users with MANAGER or OWNER role at the adjustment’s location.
- The manager reviews the adjustment request. They can:
- Approve: The adjustment is applied to inventory. The status changes to
APPROVED. Inventory quantity is updated. A movement record is logged. - Reject: The adjustment is not applied. The status changes to
REJECTED. The requesting staff member is notified with the rejection reason. Inventory is unchanged.
- Approve: The adjustment is applied to inventory. The status changes to
- Rejected adjustments can be revised and resubmitted as new adjustment requests.
Adjustment Approval Sequence
sequenceDiagram
autonumber
participant S as Staff
participant UI as Inventory UI
participant API as Backend
participant DB as DB
participant NOTIF as Notification Service
participant M as Manager
S->>UI: Identify Discrepancy
S->>UI: Click "New Adjustment"
S->>UI: Select Product, Location
S->>UI: Enter Qty Change (+3 or -2)
S->>UI: Select Reason Code
S->>UI: Enter Notes
UI->>API: POST /adjustments
API->>DB: Create Adjustment (Status: PENDING)
Note right of DB: ADJ-2026-00015 created
Note right of DB: Inventory NOT changed yet
API->>NOTIF: Send Approval Request to Manager(s)
API-->>UI: "Adjustment submitted for approval"
NOTIF-->>M: "Adjustment ADJ-2026-00015: +3 units of SKU NXJ1078 at Store A — awaiting approval"
alt Manager Approves
M->>UI: Review Adjustment -> Click "Approve"
UI->>API: POST /adjustments/{id}/approve
API->>DB: Update Status: APPROVED
API->>DB: Update Inventory Qty (apply qty_change)
API->>DB: Record approved_by, approved_at
API->>DB: Log Movement Record (ADJUSTMENT_UP or ADJUSTMENT_DOWN)
API->>NOTIF: Notify Staff: "Adjustment approved"
NOTIF-->>S: "Your adjustment ADJ-2026-00015 was approved"
API-->>UI: Adjustment Approved — Inventory Updated
else Manager Rejects
M->>UI: Review Adjustment -> Click "Reject"
M->>UI: Enter Rejection Reason
UI->>API: POST /adjustments/{id}/reject
API->>DB: Update Status: REJECTED
API->>DB: Record rejection_reason
API->>NOTIF: Notify Staff: "Adjustment rejected"
NOTIF-->>S: "Your adjustment ADJ-2026-00015 was rejected: reason"
API-->>UI: Adjustment Rejected — Inventory Unchanged
end
4.7.2 Adjustment Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
adjustment_number | String | Yes | Auto-generated: ADJ-{YEAR}-{SEQ} |
product_id | UUID | Yes | Reference to product |
variant_id | UUID | No | Reference to specific variant (if applicable) |
location_id | UUID | Yes | Location where adjustment applies |
qty_change | Integer | Yes | Positive (found stock) or negative (shrinkage/damage). Must not be 0. |
reason_code | String | Yes | Standard reason code or custom reason code (see Section 4.7.3). |
notes | Text | No | Explanation of adjustment. Mandatory for certain reason codes (e.g., OTHER). |
requested_by | UUID | Yes | Staff member who requested the adjustment |
approved_by | UUID | No | Manager who approved (null until approved) |
rejected_by | UUID | No | Manager who rejected (null unless rejected) |
approval_status | Enum | Yes | PENDING, APPROVED, REJECTED |
rejection_reason | Text | No | Manager’s reason for rejection (if rejected) |
cost_impact | Decimal(10,2) | Computed | qty_change x weighted_avg_cost. Positive for found stock, negative for shrinkage. Calculated at approval time. |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
approved_at | DateTime | No | Timestamp of approval |
rejected_at | DateTime | No | Timestamp of rejection |
4.7.3 Custom Reason Codes
The system provides a standard set of reason codes for inventory adjustments. In addition, tenants can define custom reason codes to capture business-specific adjustment scenarios.
Standard Reason Codes
| Code | Direction | Description |
|---|---|---|
DAMAGED | Negative | Item identified as damaged and removed from sellable inventory. |
THEFT | Negative | Item believed to be stolen (shoplifting, employee theft). |
COUNT_CORRECTION | Either | Correction resulting from a stock count variance that was not captured in the count approval process. |
SAMPLE | Negative | Item removed from inventory and given as a sample (to vendor, customer, or for marketing). |
WRITE_OFF | Negative | Item permanently removed from inventory as a loss. Requires cost documentation for accounting. |
FOUND_STOCK | Positive | Stock discovered that was not in the system (e.g., found behind a shelf, unscanned box). |
RETURN_TO_STOCK | Positive | Item returned to inventory outside the normal return-to-stock workflow (e.g., item used for display purposes and now returned to sellable stock). |
OTHER | Either | None of the above. Notes field becomes mandatory. |
Custom Reason Code Data Model
Tenants can create additional reason codes beyond the standard set. Custom reason codes behave identically to standard codes in all workflows – they appear in the reason code dropdown, are logged in movement records, and are available in reports.
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
code | String(50) | Yes | Unique code identifier (e.g., EMPLOYEE_PURCHASE, PHOTO_SHOOT, CHARITY_DONATION). Must be uppercase, alphanumeric with underscores. Must not conflict with standard reason codes. |
display_name | String(100) | Yes | Human-readable name shown in the UI dropdown (e.g., “Employee Purchase”, “Photo Shoot”, “Charity Donation”). |
description | Text | No | Optional description of when this reason code should be used. |
direction | Enum | Yes | POSITIVE, NEGATIVE, or BOTH. Controls whether this code can be used for positive adjustments, negative adjustments, or both. |
requires_notes | Boolean | Yes | Whether the notes field is mandatory when this reason code is selected. Default: false. |
is_active | Boolean | Yes | Whether this custom reason code is available for selection. Inactive codes are hidden from the dropdown but preserved in historical records. |
sort_order | Integer | Yes | Display order in the reason code dropdown (after standard codes). |
tenant_id | UUID | Yes | Owning tenant |
created_by | UUID | Yes | User who created the custom reason code |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Custom reason codes are tenant-specific. Each tenant manages their own set of custom codes.
- Custom code identifiers must be unique within a tenant and must not duplicate any standard code.
- Custom codes can be deactivated (soft delete) but not hard-deleted, since historical adjustment records may reference them.
- Only users with MANAGER or OWNER role can create, edit, or deactivate custom reason codes.
- When a custom reason code has
requires_notes = true, the adjustment form enforces a mandatory notes field when that code is selected. - Custom reason codes appear in all reports alongside standard codes. Reports can filter by standard vs. custom codes.
- Standard reason codes cannot be deactivated or modified by tenants. They are system-level constants.
4.7.4 Business Rules Summary
- All adjustments require manager approval. Positive adjustments (found stock), negative adjustments (shrinkage, damage), and zero-net adjustments (reclassification) all require explicit manager approval before inventory quantities change. There is no auto-approval threshold.
- Adjustment status flow:
PENDING(created, awaiting approval) ->APPROVED(manager approved, inventory updated) orREJECTED(manager rejected, inventory unchanged). - Inventory is not changed until approval. The
PENDINGadjustment is a request only. The physical inventory quantity in the system remains unchanged until a manager explicitly approves the adjustment. WRITE_OFFadjustments require cost documentation. When the reason code isWRITE_OFF, the system calculates and records thecost_impactfield for accounting reconciliation. The manager can see the dollar impact before approving.- All approved adjustments are logged as
ADJUSTMENT_UP(positive qty_change) orADJUSTMENT_DOWN(negative qty_change) movements in the movement history audit trail. - Rejected adjustments are preserved. Rejected adjustments remain in the system for audit purposes. They are not deleted. The rejection reason is recorded.
- Concurrent adjustment protection: If two staff members submit adjustments for the same product at the same location, the manager sees both pending adjustments and can approve or reject each independently. The system recalculates the inventory impact at approval time based on the current quantity, not the quantity at request time.
4.7.5 Reports: Inventory Adjustments
| Report | Purpose | Key Data Fields |
|---|---|---|
| Adjustment History | All manual inventory adjustments | Adjustment number, product, location, qty change, reason code (standard/custom), requested by, approved/rejected by, date, cost impact |
| Pending Adjustments | Adjustments awaiting manager approval | Adjustment number, product, location, qty change, reason code, requested by, request date, days pending |
| Shrinkage Report | Track inventory loss by reason code and location | Period, location, reason code, qty lost, value at cost, % of total inventory value |
| Reason Code Analysis | Frequency and impact of each reason code | Reason code, adjustment count, total qty impact, total cost impact, avg approval time, rejection rate |
| Custom Reason Code Usage | Track usage of tenant-defined reason codes | Custom code, display name, adjustment count, total qty impact, last used date |
4.8 Inter-Store Transfers
Scope: Moving inventory between store locations and HQ warehouse with full workflow tracking, variance detection on receipt, and auto-rebalancing recommendations based on sales velocity analysis. This section covers bi-directional transfer initiation (HQ push and store pull), manual allocation for scarce items, and auto-suggest workflows. The customer-facing paid transfer request (from Section 1.7) feeds into this system as the triggering event.
4.8.1 Transfer State Machine
The transfer lifecycle supports 10 states covering the full journey from request through completion, including rejection and cancellation paths.
stateDiagram-v2
[*] --> REQUESTED: Transfer Requested
REQUESTED --> APPROVED: Source Manager Approves
REQUESTED --> REJECTED: Source Manager Rejects
APPROVED --> PICKING: Pick List Generated
PICKING --> SHIPPED: Items Shipped
SHIPPED --> IN_TRANSIT: Carrier Confirmed Pickup
IN_TRANSIT --> RECEIVED: Destination Receives
RECEIVED --> COMPLETED: All Items Verified
APPROVED --> CANCELLED: Cancelled After Approval
REJECTED --> CLOSED: No Further Action
CANCELLED --> CLOSED: No Further Action
note right of REQUESTED
Destination store or HQ requests stock
OR HQ pushes stock to store
Awaiting source approval
end note
note right of APPROVED
Source manager authorized the transfer
Pick list ready for warehouse/store
end note
note right of PICKING
Source store picking items
Inventory not yet decremented
end note
note right of SHIPPED
Items handed to carrier or internal transport
Source inventory decremented
Tracking number recorded
end note
note right of IN_TRANSIT
Carrier confirmed pickup
Items between locations
end note
note right of RECEIVED
Destination received and is verifying
Variances being recorded
end note
note right of COMPLETED
All items accounted for
Destination inventory incremented
end note
4.8.2 Transfer Initiation Directions
Transfers can be initiated in two directions. Both directions enter the state machine at the same REQUESTED state.
Pull Model (Store Requests from Source)
A destination store identifies a need (low stock, customer demand, auto-suggest alert) and creates a transfer request specifying the source location and desired items. The source manager reviews and approves or rejects.
Use cases:
- Store manager sees low stock on a bestseller and requests units from HQ or another store.
- Customer requests an item available at another location (paid transfer from Section 1.7).
- System auto-suggest identifies an imbalance and the destination manager initiates the transfer.
Push Model (HQ Pushes to Stores)
HQ warehouse staff or a regional manager initiates a transfer from HQ to one or more stores. The HQ manager acts as both requester and approver, so the transfer can skip the approval wait if the same user has both roles.
Use cases:
- New seasonal inventory arrives at HQ and is distributed to stores.
- HQ manager reviews rebalancing suggestions and pushes stock to understocked stores.
- Vendor replacement shipment received at HQ needs distribution to affected stores.
Initiation Direction Data
The transfer header includes a direction field to distinguish the two models:
| Field | Type | Required | Description |
|---|---|---|---|
direction | Enum | Yes | PULL (destination requests) or PUSH (source initiates) |
initiated_by_location_id | UUID | Yes | Location that created the transfer (source for PUSH, destination for PULL) |
Business Rules for Direction:
- PULL transfers require source manager approval before proceeding to PICKING.
- PUSH transfers from HQ may be auto-approved if the initiating user holds the
inventory_manageroradminrole at the source location. - PUSH transfers between peer stores (non-HQ) still require source manager approval.
- Both directions produce identical downstream workflow (PICKING through COMPLETED).
4.8.3 Transfer Data Model
Transfer Header
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
transfer_number | String | Yes | Auto-generated: TRF-{YEAR}-{SEQ} (e.g., TRF-2026-00088) |
source_location_id | UUID | Yes | Location sending stock |
destination_location_id | UUID | Yes | Location receiving stock |
direction | Enum | Yes | PULL (destination requests) or PUSH (source initiates) |
initiated_by_location_id | UUID | Yes | Location that created the transfer |
status | Enum | Yes | REQUESTED, APPROVED, REJECTED, PICKING, SHIPPED, IN_TRANSIT, RECEIVED, COMPLETED, CANCELLED, CLOSED |
priority | Enum | No | NORMAL, URGENT, CUSTOMER_REQUEST (default: NORMAL) |
requested_by | UUID | Yes | Staff member who initiated the request |
approved_by | UUID | No | Manager who approved/rejected |
shipped_date | Date | No | Date items were shipped |
received_date | Date | No | Date items were received |
tracking_number | String | No | Carrier tracking number |
carrier | String | No | Carrier name (e.g., internal, UPS, FedEx) |
estimated_arrival | Date | No | Expected delivery date |
auto_suggest_id | UUID | No | FK to auto-suggest recommendation that triggered this transfer (null if manually created) |
notes | Text | No | Transfer notes |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Transfer Line Items
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
transfer_id | UUID | Yes | Reference to parent transfer |
product_id | UUID | Yes | Product being transferred |
variant_id | UUID | No | Specific variant (if applicable) |
qty_requested | Integer | Yes | Quantity requested by destination |
qty_shipped | Integer | No | Actual quantity shipped by source |
qty_received | Integer | No | Quantity verified at destination |
variance | Integer | Computed | Calculated: qty_received - qty_shipped |
variance_notes | Text | No | Explanation of any variance |
condition_on_receive | Enum | No | GOOD, DAMAGED, WRONG_ITEM |
4.8.4 Transfer Workflow (Pull Model)
sequenceDiagram
autonumber
participant DST as Destination Staff
participant SRC_M as Source Manager
participant SRC as Source Staff
participant UI as Transfer UI
participant API as Backend
participant DB as DB
Note over DST, DB: Step 1: Request Transfer (Pull)
DST->>UI: Click "Request Transfer"
DST->>UI: Select Source Location
DST->>UI: Add Products & Quantities
UI->>API: POST /transfers
API->>DB: Create Transfer (Status: REQUESTED, Direction: PULL)
API-->>DST: Transfer #TRF-2026-00088 Created
API-->>SRC_M: Notification: "Transfer request from Store B"
Note over SRC_M, DB: Step 2: Approve/Reject
SRC_M->>UI: Review Transfer Request
SRC_M->>UI: Check Source Stock Availability
alt Approve
SRC_M->>UI: Click "Approve"
UI->>API: POST /transfers/{id}/approve
API->>DB: Update Status: APPROVED
API->>DB: Generate Pick List
else Reject
SRC_M->>UI: Click "Reject" + Enter Reason
UI->>API: POST /transfers/{id}/reject
API->>DB: Update Status: REJECTED -> CLOSED
end
Note over SRC, DB: Step 3: Pick & Ship
SRC->>UI: Open Pick List
SRC->>UI: Pick Items (Scan to Verify)
SRC->>UI: Enter Qty Shipped per Line
SRC->>UI: Enter Tracking Number (if applicable)
SRC->>UI: Click "Ship"
UI->>API: POST /transfers/{id}/ship
API->>DB: Decrement Source Inventory
API->>DB: Update Status: SHIPPED
API->>DB: Log TRANSFER_OUT Movement (Section 4.12)
Note over DST, DB: Step 4: Receive & Verify
DST->>UI: Open Transfer -> Click "Receive"
loop Verify Each Line Item
DST->>UI: Scan/Count Received Items
DST->>UI: Enter Qty Received per Line
opt Variance
UI-->>DST: "Shipped: 20, Received: 18"
DST->>UI: Enter Variance Notes
end
opt Damaged Items
DST->>UI: Mark Condition: DAMAGED
end
end
DST->>UI: Click "Confirm Receive"
UI->>API: POST /transfers/{id}/receive
par Post-Receive
API->>DB: Increment Destination Inventory
API->>DB: Log TRANSFER_IN Movement (Section 4.12)
API->>DB: Record Variances
API->>DB: Update Status: COMPLETED
end
API-->>DST: Transfer Complete
4.8.6 Transfer Workflow (Push Model)
sequenceDiagram
autonumber
participant HQ as HQ Manager
participant SRC as HQ Warehouse Staff
participant UI as Transfer UI
participant API as Backend
participant DB as DB
participant DST as Destination Staff
Note over HQ, DST: Step 1: HQ Initiates Push Transfer
HQ->>UI: Click "Push Stock to Store"
HQ->>UI: Select Destination Store(s)
HQ->>UI: Add Products & Quantities
UI->>API: POST /transfers
API->>DB: Create Transfer (Status: REQUESTED, Direction: PUSH)
alt HQ Manager Has Approval Authority
API->>DB: Auto-Approve (Status: APPROVED)
API->>DB: Generate Pick List
API-->>HQ: Transfer Auto-Approved, Pick List Ready
else Requires Separate Approval
API-->>HQ: Transfer Created, Awaiting Approval
end
Note over SRC, DB: Step 2: Pick & Ship (same as Pull model)
SRC->>UI: Open Pick List
SRC->>UI: Pick Items (Scan to Verify)
SRC->>UI: Enter Qty Shipped per Line
SRC->>UI: Enter Tracking Number
SRC->>UI: Click "Ship"
UI->>API: POST /transfers/{id}/ship
API->>DB: Decrement Source Inventory
API->>DB: Update Status: SHIPPED
API->>DB: Log TRANSFER_OUT Movement (Section 4.12)
API-->>DST: Notification: "Incoming transfer from HQ"
Note over DST, DB: Step 3: Receive & Verify (same as Pull model)
DST->>UI: Open Transfer -> Click "Receive"
DST->>UI: Verify Items, Enter Qty Received
UI->>API: POST /transfers/{id}/receive
API->>DB: Increment Destination Inventory
API->>DB: Log TRANSFER_IN Movement (Section 4.12)
API->>DB: Update Status: COMPLETED
API-->>DST: Transfer Complete
4.8.7 Auto-Suggest Transfers
The system continuously monitors inventory distribution relative to sales velocity across all locations and generates transfer suggestions when significant imbalances are detected.
Auto-Suggest Algorithm
Step 1: Calculate Days of Supply per Product per Location
days_of_supply = qty_on_hand / avg_daily_velocity
Where avg_daily_velocity is the trailing 30-day sales average (configurable per tenant).
Step 2: Detect Imbalances
An imbalance is flagged when:
- One location has > 60 days of supply (overstocked threshold, configurable)
- Another location has < 15 days of supply (understocked threshold, configurable)
- Both locations are active retail stores (HQ warehouse uses separate thresholds)
Step 3: Calculate Suggested Quantity
target_days_of_supply = 30 (configurable per tenant)
qty_needed = (target_days_of_supply - current_days_of_supply) x avg_daily_velocity
qty_available_to_send = qty_on_hand - (target_days_of_supply x avg_daily_velocity)
suggested_qty = MIN(qty_needed at destination, qty_available_to_send from source)
The algorithm ensures the source location retains at least target_days_of_supply worth of stock after the transfer.
Step 4: Generate Suggestion
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
product_id | UUID | Yes | Product with detected imbalance |
variant_id | UUID | No | Specific variant (if applicable) |
source_location_id | UUID | Yes | Overstocked location (sender) |
destination_location_id | UUID | Yes | Understocked location (receiver) |
suggested_qty | Integer | Yes | Recommended transfer quantity |
source_days_of_supply | Decimal(8,1) | Yes | Current days of supply at source |
destination_days_of_supply | Decimal(8,1) | Yes | Current days of supply at destination |
source_velocity | Decimal(8,2) | Yes | Avg daily sales at source |
destination_velocity | Decimal(8,2) | Yes | Avg daily sales at destination |
status | Enum | Yes | PENDING, APPROVED, REJECTED, EXPIRED, CONVERTED |
reviewed_by | UUID | No | Manager who reviewed the suggestion |
reviewed_at | DateTime | No | Timestamp of review |
transfer_id | UUID | No | FK to transfer created from this suggestion (if approved) |
batch_id | UUID | Yes | Groups suggestions from the same analysis run |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Timestamp of suggestion generation |
Auto-Suggest Workflow
sequenceDiagram
autonumber
participant CRON as Scheduler (Weekly/On-Demand)
participant SVC as Rebalancing Service
participant DB as DB
participant MGR as Manager
participant UI as Dashboard UI
participant API as Backend
Note over CRON, API: Phase 1: Generate Suggestions
CRON->>SVC: Trigger Rebalancing Analysis
SVC->>DB: Query qty_on_hand per product per location
SVC->>DB: Query trailing 30-day sales velocity per product per location
SVC->>SVC: Calculate days_of_supply for each product/location
SVC->>SVC: Detect imbalances (>60 days vs <15 days)
SVC->>SVC: Calculate suggested transfer quantities
SVC->>DB: Insert suggestions (Status: PENDING, batch_id: {batch})
SVC-->>MGR: Notification: "12 rebalancing suggestions ready for review"
Note over MGR, API: Phase 2: Manager Review
MGR->>UI: Open Rebalancing Dashboard
UI->>API: GET /transfer-suggestions?status=PENDING
API-->>UI: Return suggestion list with velocity data
loop Review Each Suggestion
UI-->>MGR: Show: Product, Source (85 days supply), Dest (8 days supply), Suggested Qty: 15
alt Approve Suggestion
MGR->>UI: Click "Approve" (may adjust qty)
UI->>API: POST /transfer-suggestions/{id}/approve
API->>DB: Update suggestion status: APPROVED
API->>DB: Create Transfer (Status: REQUESTED, auto_suggest_id: {suggestion_id})
API-->>MGR: Transfer #TRF-2026-00090 Created
else Reject Suggestion
MGR->>UI: Click "Reject" + Enter Reason
UI->>API: POST /transfer-suggestions/{id}/reject
API->>DB: Update suggestion status: REJECTED
end
end
Note over MGR, API: Phase 3: Approve All (Batch)
opt Batch Approve
MGR->>UI: Click "Approve All Remaining"
UI->>API: POST /transfer-suggestions/batch/{batch_id}/approve
API->>DB: Update all PENDING to APPROVED
API->>DB: Create Transfer records for each
API-->>MGR: "8 transfers created from suggestions"
end
Note over SVC, DB: Phase 4: Expiration
SVC->>DB: Expire PENDING suggestions older than 7 days
SVC->>DB: Update status: EXPIRED
Configuration:
| Setting | Default | Description |
|---|---|---|
rebalance_schedule | Weekly (Monday 6:00 AM) | When auto-suggest analysis runs |
overstocked_threshold_days | 60 | Days of supply above which a location is considered overstocked |
understocked_threshold_days | 15 | Days of supply below which a location is considered understocked |
target_days_of_supply | 30 | Target days of supply after rebalancing |
velocity_lookback_days | 30 | Trailing days used to calculate average daily velocity |
suggestion_expiry_days | 7 | Days before unreviewed suggestions expire |
min_suggested_qty | 1 | Minimum quantity for a suggestion to be generated |
hq_overstocked_threshold_days | 90 | Separate overstocked threshold for HQ warehouse |
Business Rules:
- Auto-suggest never creates transfers automatically. All suggestions require manager review.
- Manager can modify the suggested quantity before approving (e.g., reduce from 15 to 10).
- Suggestions expire after
suggestion_expiry_daysif not reviewed. Expired suggestions are excluded from the next run to avoid duplicates. - The algorithm excludes products with zero velocity at both source and destination (dead stock requires manual review, not rebalancing).
- HQ warehouse uses separate thresholds because HQ holds distribution stock, not retail selling stock.
- If multiple destinations need the same product from the same source, the system generates one suggestion per source-destination pair.
4.8.8 Manual Allocation for Scarce Items
When multiple stores need the same scarce item and HQ has limited stock, the system does not attempt to automatically split the available quantity. Instead, a manager manually decides the allocation based on business judgment.
Scarce Item Scenario
A scarce item condition exists when:
- Two or more stores have submitted transfer requests (or auto-suggest has flagged multiple destinations) for the same product.
- The source location (typically HQ) does not have enough stock to fulfill all requests in full.
Allocation Dashboard
When a scarce item condition is detected, the system presents an Allocation Dashboard that consolidates all competing requests:
| Column | Description |
|---|---|
| Product | SKU, name, variant |
| HQ Available Qty | Current on-hand at HQ (source) |
| Store | Each requesting store listed as a row |
| Requested Qty | Quantity each store is requesting |
| Current On-Hand | Quantity each store currently holds |
| Days of Supply | Calculated days of supply at each store |
| 30-Day Velocity | Average daily sales at each store |
| Allocated Qty | Editable field – manager enters allocation per store |
Example:
| Product | HQ Available | Store | Requested | On-Hand | Days of Supply | 30-Day Velocity | Allocated |
|---|---|---|---|---|---|---|---|
| BLK-TEE-M | 20 | Store GM | 12 | 2 | 4 days | 0.5/day | ___ |
| BLK-TEE-M | 20 | Store HM | 15 | 1 | 2 days | 0.5/day | ___ |
| BLK-TEE-M | 20 | Store NM | 8 | 5 | 12 days | 0.4/day | ___ |
| Total Requested | 35 | ___ / 20 |
The manager sees that total requests (35) exceed available stock (20) and enters allocations that sum to at most 20. The system validates that SUM(allocated_qty) <= available_qty.
Business Rules:
- No automated splitting. The system presents data; the human decides.
- Manager can allocate zero to any store (decline that store’s request entirely).
- The allocation creates individual transfers for each store receiving stock.
- Priority field on each transfer request (
NORMAL,URGENT,CUSTOMER_REQUEST) is displayed to help inform the manager’s decision. - Stores with
CUSTOMER_REQUESTpriority (paid customer transfers from Section 1.7) should generally be prioritized to avoid customer disappointment. - The allocation dashboard is accessible from the Transfer Management screen when the system detects competing requests for the same product.
4.8.9 Variance Handling
When the destination receives a different quantity than was shipped, a variance is recorded.
Variance Types:
| Variance | Description | Resolution |
|---|---|---|
| Short | Received less than shipped (e.g., shipped 20, received 18) | Record variance notes. Investigate: lost in transit, miscount at source, or carrier damage. Source location does not get stock back automatically; requires adjustment (Section 4.7) if items are confirmed lost. |
| Over | Received more than shipped (e.g., shipped 20, received 22) | Rare. Likely miscount at source. Record variance. Destination inventory reflects actual received count. |
| Damaged | Items received in damaged condition | Record condition as DAMAGED. Damaged items enter damaged inventory status at destination. May trigger RMA (Section 4.9) or write-off. |
| Wrong Item | Different product received than expected | Record condition as WRONG_ITEM. Requires follow-up: return to source or create adjustment at both locations. |
Business Rules:
- Variance percentage > 10% triggers a notification to both source and destination managers.
- All variances are logged in the Product Movement History (Section 4.12) with reason codes.
- Unresolved variances appear on the Transfer Variance Report for management review.
4.8.10 Carrier Tracking
For transfers shipped via external carriers (not internal transport), tracking information is recorded.
| Field | Type | Required | Description |
|---|---|---|---|
carrier | String | No | Carrier name: INTERNAL, UPS, FEDEX, USPS, DHL, OTHER |
tracking_number | String | No | Carrier tracking number |
estimated_arrival | Date | No | Expected delivery date |
actual_arrival | Date | No | Actual delivery date (set on receive) |
shipping_cost | Decimal(10,2) | No | Cost of shipment (for internal cost tracking) |
Business Rules:
- Carrier and tracking number are required when
carrieris notINTERNAL. INTERNALcarrier indicates the transfer is hand-delivered by staff or via company vehicle.- The system does not integrate with carrier tracking APIs in v1. Tracking numbers are recorded for manual lookup.
4.8.11 Reports: Transfers
| Report | Purpose | Key Data Fields |
|---|---|---|
| Open Transfer Report | Track in-progress transfers | Transfer number, source, destination, direction (Push/Pull), status, item count, total units, days in transit, priority |
| Transfer Variance Report | Discrepancies between shipped and received | Transfer number, product, qty shipped, qty received, variance, variance %, condition, notes |
| Transfer Volume Report | Volume of transfers between locations | Source, destination, transfer count, total units transferred, total value, period, direction breakdown |
| Rebalancing Suggestions | Auto-generated transfer recommendations | Product, source location (days of supply), destination location (days of supply), suggested qty, current velocity data, suggestion status |
| Scarce Item Allocation Log | Audit trail of manual allocation decisions | Product, HQ available qty, stores requesting, qty allocated per store, manager who allocated, date |
| Transfer Lead Time | Average time from request to completion | Source, destination, carrier, avg days requested-to-shipped, avg days shipped-to-received, avg total lead time |
4.9 Vendor RMA & Returns
Scope: Managing the return of defective, damaged, incorrect, or overstock merchandise to vendors. The RMA (Return Merchandise Authorization) workflow tracks the complete lifecycle from initial request through vendor approval, shipment back to the vendor, and receipt of credit or replacement inventory. This section covers both defective/quality RMA returns and overstock returns, which follow distinct workflows.
4.9.1 RMA State Machine
stateDiagram-v2
[*] --> DRAFT: RMA Created
DRAFT --> SUBMITTED: Submit to Vendor
SUBMITTED --> VENDOR_APPROVED: Vendor Approves Return
SUBMITTED --> VENDOR_REJECTED: Vendor Rejects Return
VENDOR_APPROVED --> SHIPPED_BACK: Items Shipped to Vendor
SHIPPED_BACK --> CREDIT_RECEIVED: Vendor Issues Credit Memo
SHIPPED_BACK --> REPLACEMENT_RECEIVED: Vendor Sends Replacement
CREDIT_RECEIVED --> CLOSED: RMA Finalized
REPLACEMENT_RECEIVED --> CLOSED: RMA Finalized
VENDOR_REJECTED --> CLOSED: RMA Closed (No Action)
note right of DRAFT
Staff assembles return list
Line items editable
No inventory impact
end note
note right of SUBMITTED
Sent to vendor for review
Awaiting vendor response
Line items locked
end note
note right of VENDOR_APPROVED
Vendor authorized return
Ready to ship back
end note
note right of SHIPPED_BACK
Items in transit to vendor
Inventory decremented at source
Tracking number recorded
end note
note right of CLOSED
Credit applied or replacement received
Audit trail complete
end note
4.9.2 RMA Data Model
RMA Header
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
rma_number | String | Yes | Auto-generated: RMA-{YEAR}-{SEQ} (e.g., RMA-2026-00015) |
vendor_id | UUID | Yes | Reference to vendor |
source_location_id | UUID | Yes | Location from which items are being returned |
status | Enum | Yes | DRAFT, SUBMITTED, VENDOR_APPROVED, VENDOR_REJECTED, SHIPPED_BACK, CREDIT_RECEIVED, REPLACEMENT_RECEIVED, CLOSED |
reason | Enum | Yes | DEFECTIVE, DAMAGED, WRONG_ITEM, OVERSTOCK |
rma_type | Enum | Yes | DEFECTIVE_RETURN or OVERSTOCK_RETURN (see Section 4.9.6) |
notes | Text | No | Free-form notes about the return |
vendor_agreement_ref | String | No | Reference to vendor agreement or pre-authorization (required for OVERSTOCK returns) |
created_by | UUID | Yes | Staff member who created the RMA |
approved_by | UUID | No | Vendor contact or reference who approved |
ship_date | Date | No | Date items were shipped back to vendor |
tracking_number | String | No | Carrier tracking number for return shipment |
credit_amount | Decimal(10,2) | No | Credit amount issued by vendor |
restocking_fee_pct | Decimal(5,2) | No | Vendor restocking fee percentage (applicable to OVERSTOCK returns) |
restocking_fee_amount | Decimal(10,2) | No | Calculated restocking fee deducted from credit |
net_credit_amount | Decimal(10,2) | No | Calculated: credit_amount - restocking_fee_amount |
replacement_po_id | UUID | No | FK to replacement purchase order (if vendor sends replacement) |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
RMA Line Items
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
rma_id | UUID | Yes | Reference to parent RMA |
product_id | UUID | Yes | Reference to product being returned |
variant_id | UUID | No | Reference to specific variant (if applicable) |
qty | Integer | Yes | Quantity being returned to vendor |
unit_cost | Decimal(10,2) | Yes | Cost per unit (from original PO or weighted avg cost) |
line_total | Decimal(10,2) | Yes | Calculated: qty x unit_cost |
condition_notes | Text | No | Description of item condition |
inspection_result | Enum | No | CONFIRMED_DEFECTIVE, COSMETIC_DAMAGE, NOT_AS_DESCRIBED, NOT_INSPECTED (for overstock items) |
original_po_id | UUID | No | Reference to the original purchase order that delivered this item (for traceability) |
4.9.3 RMA Workflow (Defective Returns)
sequenceDiagram
autonumber
participant U as Staff
participant UI as RMA UI
participant API as Backend
participant DB as DB
participant V as Vendor
Note over U, V: Step 1: Create RMA
U->>UI: Click "New Vendor RMA"
UI->>UI: Select Vendor
UI->>API: GET /vendors/{id}/products
API-->>UI: Return Vendor's Products
loop Add RMA Line Items
U->>UI: Select Product
U->>UI: Enter Quantity to Return
U->>UI: Select Reason (Defective/Damaged/Wrong Item)
U->>UI: Enter Condition Notes
U->>UI: Record Inspection Result
end
U->>UI: Click "Save Draft"
UI->>API: POST /vendor-rma
API->>DB: Create RMA Record (Status: DRAFT, Type: DEFECTIVE_RETURN)
API-->>UI: RMA #RMA-2026-00015 Created
Note over U, V: Step 2: Submit to Vendor
U->>UI: Review RMA -> Click "Submit"
UI->>API: POST /vendor-rma/{id}/submit
alt Email Submission
API->>V: Send RMA via Email (PDF attachment)
else Manual Submission
API-->>UI: "RMA marked Submitted - contact vendor manually"
end
API->>DB: Update Status: SUBMITTED
API-->>UI: RMA Submitted
Note over V, U: Step 3: Vendor Response
alt Vendor Approves
U->>UI: Mark RMA as "Vendor Approved"
UI->>API: POST /vendor-rma/{id}/approve
API->>DB: Update Status: VENDOR_APPROVED
else Vendor Rejects
U->>UI: Mark RMA as "Vendor Rejected"
U->>UI: Enter Rejection Reason
UI->>API: POST /vendor-rma/{id}/reject
API->>DB: Update Status: VENDOR_REJECTED
API->>DB: Update Status: CLOSED
end
Note over U, V: Step 4: Ship Items Back
U->>UI: Enter Tracking Number & Ship Date
UI->>API: POST /vendor-rma/{id}/ship
API->>DB: Decrement Inventory at Source Location
API->>DB: Log RMA_OUT Movement (Section 4.12)
API->>DB: Update Status: SHIPPED_BACK
API-->>UI: Items Shipped
Note over V, DB: Step 5: Vendor Resolution
alt Credit Memo
V-->>U: Vendor Issues Credit Memo
U->>UI: Enter Credit Amount
UI->>API: POST /vendor-rma/{id}/credit
API->>DB: Record Credit Amount
API->>DB: Update Vendor Account Balance
API->>DB: Update Status: CREDIT_RECEIVED
else Replacement Shipment
V-->>U: Vendor Sends Replacement
U->>UI: Click "Receive Replacement"
UI->>API: POST /vendor-rma/{id}/replacement
API->>DB: Create Linked Purchase Order
API->>DB: Increment Inventory (Replacement Items)
API->>DB: Log RMA_IN Movement (Section 4.12)
API->>DB: Update Status: REPLACEMENT_RECEIVED
end
Note over U, DB: Step 6: Close RMA
U->>UI: Click "Close RMA"
UI->>API: POST /vendor-rma/{id}/close
API->>DB: Update Status: CLOSED
API-->>UI: RMA Closed
4.9.4 Business Rules (Defective Returns)
- RMA can only be created for products with an active vendor relationship (exists in
vendor_producttable with anACTIVEvendor). - Items must be in
AVAILABLEorDAMAGEDinventory status to be placed on an RMA. Items inIN_TRANSIT,RESERVED, orQUARANTINEcannot be returned until their status resolves. - Inventory is decremented from the source location when the RMA status changes to
SHIPPED_BACK, not before. This ensures accurate on-hand counts until items physically leave. - Credit amounts are reconciled with the vendor’s account balance. If the tenant tracks payables, the credit reduces the outstanding amount owed to the vendor.
- Replacement purchase orders link back to the original RMA via
replacement_po_idfor complete audit trail. - Auto-increment RMA number per tenant:
RMA-{YEAR}-{SEQUENCE}. - Each RMA line item must have an
inspection_resultrecorded before the RMA can be submitted. This ensures quality documentation accompanies the return request. - All inventory movements (out for RMA, in for replacement) are recorded in the Product Movement History (Section 4.12) with event types
RMA_OUTandRMA_IN.
4.9.5 Reports: Vendor RMA
| Report | Purpose | Key Data Fields |
|---|---|---|
| Open RMA Report | Track outstanding vendor returns | RMA number, vendor, status, rma_type, total value, days open, last action date |
| Vendor Return Rate | Quality tracking per vendor | Vendor, RMA count, units returned, return % of total purchased, top reasons, defective vs overstock breakdown |
| RMA Aging | Identify stalled returns | RMA number, vendor, current status, days in current status, last action, escalation flag |
| RMA Credit Reconciliation | Track credits received vs expected | RMA number, vendor, expected credit, actual credit, restocking fees, net credit, variance, reconciliation status |
| Overstock Return Summary | Track overstock-specific returns | Vendor, period, units returned as overstock, gross credit, restocking fees paid, net credit received |
4.9.6 Overstock Return Workflow
Overstock returns represent a fundamentally different business process from defective RMAs. Overstock returns involve negotiated return of unsold seasonal, end-of-line, or slow-moving merchandise to the vendor. The vendor has pre-agreed to accept the return, often with a restocking fee.
Key Differences from Defective RMA
| Aspect | Defective RMA | Overstock Return |
|---|---|---|
| Trigger | Quality issue discovered in stock | Excess inventory / end of season |
| Inspection | Required – each item inspected for defect | Not required – items are in sellable condition |
| Vendor Pre-Agreement | Not always required (warranty claims) | Always required – must reference agreement |
| Restocking Fee | Typically none (vendor’s quality failure) | Common – vendor charges 10-25% restocking fee |
| Credit Calculation | Full original cost or replacement | Negotiated – may differ from original cost |
| Urgency | High (defective items tie up shelf space) | Moderate (planned seasonal transition) |
| Reason Code | DEFECTIVE, DAMAGED, WRONG_ITEM | OVERSTOCK |
| RMA Type | DEFECTIVE_RETURN | OVERSTOCK_RETURN |
Overstock Return Process
sequenceDiagram
autonumber
participant MGR as Store/HQ Manager
participant UI as RMA UI
participant API as Backend
participant DB as DB
participant V as Vendor
Note over MGR, V: Pre-Condition: Vendor Agreement Exists
MGR->>UI: Click "New Overstock Return"
UI->>UI: Select Vendor
UI->>UI: Prompt: "Enter Vendor Agreement Reference"
MGR->>UI: Enter Agreement Ref (e.g., "Email 2026-01-15, 20% restocking agreed")
loop Add Overstock Items
MGR->>UI: Select Product (filter: in-stock, this vendor)
MGR->>UI: Enter Quantity to Return
UI-->>MGR: Show: Unit Cost, Extended Total
end
MGR->>UI: Enter Restocking Fee % (e.g., 20%)
UI->>UI: Calculate: Gross Credit, Restocking Fee, Net Credit
UI-->>MGR: "Gross: $2,500 | Restocking Fee (20%): $500 | Net Credit: $2,000"
MGR->>UI: Click "Save Draft"
UI->>API: POST /vendor-rma
API->>DB: Create RMA (Status: DRAFT, Type: OVERSTOCK_RETURN, Reason: OVERSTOCK)
API-->>UI: RMA #RMA-2026-00032 Created
Note over MGR, V: Submit, Vendor Response, Ship, Credit (same state machine)
MGR->>UI: Submit RMA
UI->>API: POST /vendor-rma/{id}/submit
API->>DB: Update Status: SUBMITTED
V-->>MGR: Vendor Confirms Acceptance
MGR->>UI: Mark Vendor Approved
UI->>API: POST /vendor-rma/{id}/approve
API->>DB: Update Status: VENDOR_APPROVED
MGR->>UI: Ship Items Back (enter tracking)
UI->>API: POST /vendor-rma/{id}/ship
API->>DB: Decrement Inventory
API->>DB: Log RMA_OUT Movement (Section 4.12)
API->>DB: Update Status: SHIPPED_BACK
V-->>MGR: Vendor Issues Credit (net of restocking fee)
MGR->>UI: Enter Credit Received
UI->>API: POST /vendor-rma/{id}/credit
API->>DB: Record credit_amount, restocking_fee_amount, net_credit_amount
API->>DB: Update Vendor Account Balance (net credit)
API->>DB: Update Status: CREDIT_RECEIVED
MGR->>UI: Close RMA
UI->>API: POST /vendor-rma/{id}/close
API->>DB: Update Status: CLOSED
Overstock Return Business Rules
- Vendor agreement required: The
vendor_agreement_reffield must be populated before an overstock RMA can be submitted. This is a free-text field documenting the pre-agreement (e.g., email reference, contract clause, verbal confirmation date). - No inspection step: Overstock items set
inspection_resulttoNOT_INSPECTED. The items are in sellable condition; they are simply excess. - Restocking fee handling:
restocking_fee_pctis entered by the user (e.g., 20.00 for 20%).restocking_fee_amountis calculated:SUM(line_totals) x (restocking_fee_pct / 100).net_credit_amountis calculated:credit_amount - restocking_fee_amount.- The vendor account balance is credited with the
net_credit_amount, not the gross amount.
- Credit may differ from cost: The
credit_amountfield on the header is the negotiated total credit from the vendor. This may be less than the sum of line item costs if the vendor negotiated a reduced rate. The system displays both the calculated cost total and the actual credit for variance tracking. - Reason code enforcement: When
rma_typeisOVERSTOCK_RETURN, thereasonfield is automatically set toOVERSTOCKand cannot be changed. - Eligibility: Only items in
AVAILABLEstatus can be placed on an overstock return.DAMAGEDitems should go through the defective RMA workflow instead. - Seasonal timing: Overstock returns are typically created in bulk at end-of-season. The system supports adding many line items (50+) to a single overstock RMA.
4.10 Serial & Lot Tracking
Scope: Tracking individual high-value items by serial number and managing batch/lot numbers for recall readiness and FIFO inventory management. Serial tracking captures the full chain of custody from receiving through sale. Lot tracking enables batch-level recall identification.
4.10.1 Serial Number Tracking
Serial tracking is enabled per product via the serial_tracked boolean flag on the product record. When enabled, serial numbers are captured at two critical points: receiving (inbound) and sale (outbound).
Serial Number Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
product_id | UUID | Yes | Reference to product |
serial_number | String | Yes | Unique serial number (unique per tenant) |
status | Enum | Yes | IN_STOCK, SOLD, RETURNED, RMA, WRITE_OFF |
location_id | UUID | No | Current location (null if sold or written off) |
received_at | DateTime | Yes | Timestamp when serial was first received |
received_via | Enum | Yes | PO_RECEIVE, TRANSFER_RECEIVE, RETURN_TO_STOCK, RMA_REPLACEMENT |
source_document_id | UUID | No | FK to the receiving source document |
sold_at | DateTime | No | Timestamp when sold |
sold_to_customer_id | UUID | No | Customer who purchased (if customer attached to sale) |
sale_order_id | UUID | No | Reference to the sale order |
tenant_id | UUID | Yes | Owning tenant |
Serial Number State Machine
stateDiagram-v2
[*] --> IN_STOCK: Received (PO/Transfer/Return)
IN_STOCK --> SOLD: Sold to Customer
SOLD --> RETURNED: Customer Returns Item
RETURNED --> IN_STOCK: Returned to Available Stock
IN_STOCK --> RMA: Sent Back to Vendor
IN_STOCK --> WRITE_OFF: Damaged Beyond Repair
RMA --> [*]: Vendor Received
WRITE_OFF --> [*]: Removed from Inventory
note right of IN_STOCK
Available for sale
Location tracked
end note
note right of SOLD
Customer association
Order reference
end note
Business Rules:
- Cannot sell a serial-tracked product at POS without scanning or entering the serial number.
- Cannot receive a serial-tracked product without assigning a serial number to each unit.
- Serial numbers are unique per tenant. Duplicate serial numbers for the same product are rejected.
- Customer association: serial -> customer link enables after-sale lookup (“Customer X purchased serial Y on date Z at location W”).
- Serial history is immutable. Status changes are appended, never overwritten.
- Serial number transfers between locations update the
location_idfield and create a movement record in the Product Movement History (Section 4.12).
4.10.2 Lot/Batch Tracking
Lot tracking is enabled per product via the lot_tracked boolean flag. Lot numbers are assigned at receiving and tracked through the sales lifecycle.
Lot Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
product_id | UUID | Yes | Reference to product |
lot_number | String | Yes | Lot/batch identifier (from vendor or manually assigned) |
qty_received | Integer | Yes | Total quantity received in this lot |
qty_on_hand | Integer | Yes | Current quantity remaining |
qty_sold | Integer | Yes | Total quantity sold from this lot |
received_date | Date | Yes | Date lot was received |
expiry_date | Date | No | Expiration date (optional; supported for future non-clothing use cases) |
source_po_id | UUID | No | Reference to purchase order that delivered this lot |
location_id | UUID | Yes | Location holding this lot |
tenant_id | UUID | Yes | Owning tenant |
Business Rules:
- FIFO enforcement: when selling lot-tracked items, the system selects from the oldest lot first (by
received_date). - Recall support: given a lot number, the system identifies all units – in stock (with location) and sold (with customer and date) – for recall action.
- Lot numbers can be entered manually or scanned from vendor packaging during receiving.
- Multiple lots of the same product can exist at the same location simultaneously.
- Lot quantities are decremented on sale and incremented on return. The return process associates the returned item back to its original lot where possible.
- Lot transfers between locations create a new lot record at the destination (same lot number, new location) and decrement the source lot’s
qty_on_hand.
4.10.3 Reports: Serial & Lot
| Report | Purpose | Key Data Fields |
|---|---|---|
| Serial Number Lookup | Find current status and full history of a serial | Serial number, product, current status, current location, customer (if sold), purchase date, receiving source |
| Lot Inventory | Stock on hand by lot | Lot number, product, qty received, qty on hand, qty sold, received date, age (days), location |
| Lot Trace (Recall) | Find all units from a specific lot | Lot number, product, units in stock (by location), units sold (customer, sale date, order number), units returned |
| Serial Warranty Lookup | Customer and purchase info for warranty claims | Serial number, product, customer name, purchase date, purchase location, order number |
4.11 Landed Cost & Costing
Scope: Tracking the true cost of inventory by calculating landed cost (purchase price plus all additional costs to get product to the selling floor) and maintaining weighted average cost across multiple purchase orders. Accurate costing is critical for margin reporting and inventory valuation.
4.11.1 Landed Cost Calculation
Landed cost captures the total acquisition cost per unit, including all expenses beyond the vendor’s unit price.
Components:
| # | Component | Description | Example |
|---|---|---|---|
| 1 | Unit Cost | Vendor price per unit (from PO line item) | $25.00 per unit |
| 2 | Freight/Shipping | Carrier charges allocated per unit | $500 total / 200 units = $2.50/unit |
| 3 | Duties/Tariffs | Import duties allocated per unit (based on product category and origin country) | $300 total / 200 units = $1.50/unit |
| 4 | Customs/Brokerage | Customs clearance fees allocated per unit | $100 total / 200 units = $0.50/unit |
| 5 | Handling | Warehouse handling fees allocated per unit | $80 total / 200 units = $0.40/unit |
Formula:
landed_cost_per_unit = unit_cost + (freight / units) + (duties / units) + (customs / units) + (handling / units)
Example: $25.00 + $2.50 + $1.50 + $0.50 + $0.40 = $29.90 landed cost per unit
Cost Allocation Methods
The system supports three methods for distributing PO-level costs across individual line items:
| Method | Allocation Logic | Best For |
|---|---|---|
| BY_UNIT | Equal cost per unit across all lines | Uniform-size items (e.g., t-shirts in poly bags) |
| BY_VALUE | Proportional to unit cost (higher-cost items absorb more) | Mixed-value POs (e.g., accessories + outerwear) |
| BY_WEIGHT | Proportional to item weight | Heavy items driving freight cost (e.g., denim vs. silk) |
Cost Allocation Data Model (per PO)
| Field | Type | Required | Description |
|---|---|---|---|
po_id | UUID | Yes | Reference to purchase order |
freight_total | Decimal(10,2) | No | Total freight cost for PO shipment |
duties_total | Decimal(10,2) | No | Total duties and tariffs |
customs_total | Decimal(10,2) | No | Customs brokerage fees |
handling_total | Decimal(10,2) | No | Warehouse handling fees |
allocation_method | Enum | Yes | BY_UNIT (equal per unit), BY_VALUE (proportional to unit cost), BY_WEIGHT |
tenant_id | UUID | Yes | Owning tenant |
Per-Line Landed Cost Data Model
| Field | Type | Required | Description |
|---|---|---|---|
po_line_id | UUID | Yes | Reference to PO line item |
unit_cost | Decimal(10,2) | Yes | Base vendor price per unit |
freight_per_unit | Decimal(10,2) | No | Allocated freight per unit |
duties_per_unit | Decimal(10,2) | No | Allocated duties per unit |
customs_per_unit | Decimal(10,2) | No | Allocated customs per unit |
handling_per_unit | Decimal(10,2) | No | Allocated handling per unit |
landed_cost_per_unit | Decimal(10,2) | Yes | Sum of all cost components |
4.11.2 Weighted Average Cost
The system maintains a weighted average cost per product per location. This cost is recalculated on every PO receive event.
Formula:
new_avg_cost = ((existing_qty x existing_avg_cost) + (received_qty x landed_cost_per_unit))
/ (existing_qty + received_qty)
Example:
- Existing: 100 units at $28.00 avg cost = $2,800.00
- Received: 50 units at $29.90 landed cost = $1,495.00
- New avg cost: ($2,800 + $1,495) / (100 + 50) = $28.63
Weighted Average Cost Data Model
| Field | Type | Required | Description |
|---|---|---|---|
product_id | UUID | Yes | Reference to product |
location_id | UUID | Yes | Reference to store/warehouse location |
weighted_avg_cost | Decimal(10,4) | Yes | Current weighted average cost (4 decimal places for precision) |
last_updated_at | DateTime | Yes | Timestamp of last recalculation |
Business Rules:
- Recalculated on every PO receive event using landed cost (not raw vendor cost).
- Used for: COGS calculation, margin reporting, inventory valuation, shrinkage valuation.
- Historical cost snapshots are preserved for each receive event to support audit and retroactive analysis.
- Initial weighted average cost for a new product is set to the first PO’s landed cost.
- Transfers between locations do not change the weighted average cost. The receiving location inherits the sending location’s cost for those units.
- Inventory adjustments (Section 4.7) use the current weighted average cost for valuation of adjusted quantities.
- Write-offs and shrinkage are valued at the weighted average cost at the time of the event.
4.11.3 Reports: Costing
| Report | Purpose | Key Data Fields |
|---|---|---|
| Landed Cost Analysis | Breakdown of cost components per PO | PO number, product, unit cost, freight, duties, customs, handling, total landed cost, allocation method |
| Margin Analysis | True margin using landed cost | Product, selling price, weighted avg cost, gross margin $, gross margin %, comparison to target margin |
| Inventory Valuation | Total inventory value at cost | Location, product count, total units, total value (at weighted avg cost), value by status |
| Cost Trend | Cost changes over time per product | Product, vendor, landed cost per PO over time, cost trend direction, % change |
4.12 Product Movement History & Stock Ledger
Scope: Maintaining a complete audit trail of every inventory movement for each product across all locations. The movement history serves as the authoritative record for inventory reconciliation, shrinkage analysis, and regulatory compliance. Every change to inventory quantity must be traced to a source document. This is the single source of truth for “what happened to inventory and why.”
4.12.1 Movement Audit Trail
Every inventory change creates a movement record. No inventory quantity changes without a corresponding movement entry.
| Event Type | Description | Qty Impact | Source Document |
|---|---|---|---|
PO_RECEIVE | Received from vendor via purchase order | +qty | Purchase Order |
TRANSFER_OUT | Shipped to another location (Section 4.8) | -qty | Transfer |
TRANSFER_IN | Received from another location (Section 4.8) | +qty | Transfer |
SALE | Sold to customer | -qty | Sale Order |
RETURN | Customer return to stock | +qty | Return Order |
ADJUSTMENT_UP | Manual positive adjustment (Section 4.7) | +qty | Adjustment |
ADJUSTMENT_DOWN | Manual negative adjustment (Section 4.7) | -qty | Adjustment |
WRITE_OFF | Inventory write-off (damaged, expired, theft) | -qty | Write-Off |
RMA_OUT | Shipped back to vendor via RMA (Section 4.9) | -qty | Vendor RMA |
RMA_IN | Replacement received from vendor via RMA (Section 4.9) | +qty | Vendor RMA |
COUNT_ADJUST | Adjustment from stock count variance | +/- qty | Stock Count |
RESERVE | Reserved for order or transfer | -available | Reservation |
UNRESERVE | Reservation released | +available | Reservation |
4.12.2 Movement Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
product_id | UUID | Yes | Reference to product |
variant_id | UUID | No | Reference to specific variant (if applicable) |
location_id | UUID | Yes | Location where movement occurred |
event_type | Enum | Yes | One of the 13 event types listed above |
qty_change | Integer | Yes | Positive (inbound) or negative (outbound) quantity change |
running_balance | Integer | Yes | Running balance at this location after this movement |
source_document_type | String | Yes | Type of source document (e.g., “PurchaseOrder”, “Transfer”, “SaleOrder”, “Adjustment”, “VendorRMA”, “StockCount”) |
source_document_id | UUID | Yes | FK to the source document |
reference_number | String | No | Human-readable reference (e.g., PO-2026-00042, TRF-2026-00088, ADJ-2026-00005) |
actor_id | UUID | Yes | User or system process that caused the movement |
reason_code | String | No | Reason code for adjustments and write-offs (see Section 4.7 for adjustment reason codes) |
notes | Text | No | Additional context |
tenant_id | UUID | Yes | Owning tenant |
created_at | DateTime | Yes | Timestamp of movement (immutable) |
Business Rules:
- Movement records are immutable. Once created, they cannot be edited or deleted. Corrections are made by creating a new, opposite movement (e.g., an
ADJUSTMENT_UPto correct an erroneousADJUSTMENT_DOWN). - The
running_balanceis computed at insert time:previous_running_balance + qty_change. This provides an instant snapshot of inventory level at any point in time without recalculating from the beginning. - Every module that changes inventory quantity (PO receiving, sales, transfers, adjustments, RMAs, stock counts) must insert a movement record as part of the same database transaction. No inventory change should be committed without its corresponding movement.
- The
actor_idfield distinguishes between human actions (staff user ID) and system actions (system process ID for automated events like auto-suggest transfers or scheduled adjustments).
4.12.3 Stock Ledger
The stock ledger provides a chronological, per-product, per-location view of all movements with running balances.
Features:
- Running balance updated with every movement. The
running_balancefield on each movement record represents the inventory level immediately after that movement was applied. - Ledger view: chronological list of all movements for a product at a location, showing date, event type, quantity change, running balance, source document, and actor.
- Drill-down: clicking any ledger entry navigates to the source document (PO, sale, transfer, adjustment, RMA, etc.) for full context.
- Reconciliation: compare the latest ledger running balance to a physical count to identify discrepancies. Any difference indicates missing or undocumented movements.
- Date-range filtering: view movements for a specific period (e.g., last 30 days, last quarter) to analyze trends.
- Event-type filtering: view only specific movement types (e.g., show only adjustments and write-offs to analyze shrinkage).
4.12.4 Stock Ledger Entity Relationship
erDiagram
PRODUCT {
UUID id PK
String sku
String name
}
LOCATION {
UUID id PK
String name
String type
}
MOVEMENT {
UUID id PK
UUID product_id FK
UUID variant_id FK
UUID location_id FK
String event_type
Integer qty_change
Integer running_balance
String source_document_type
UUID source_document_id
String reference_number
UUID actor_id
String reason_code
DateTime created_at
}
PURCHASE_ORDER {
UUID id PK
String po_number
}
TRANSFER {
UUID id PK
String transfer_number
}
SALE_ORDER {
UUID id PK
String order_number
}
ADJUSTMENT {
UUID id PK
String adjustment_number
}
VENDOR_RMA {
UUID id PK
String rma_number
}
STOCK_COUNT {
UUID id PK
String count_number
}
PRODUCT ||--o{ MOVEMENT : "has movements"
LOCATION ||--o{ MOVEMENT : "at location"
MOVEMENT }o--|| PURCHASE_ORDER : "source: PO_RECEIVE"
MOVEMENT }o--|| TRANSFER : "source: TRANSFER_IN/OUT"
MOVEMENT }o--|| SALE_ORDER : "source: SALE/RETURN"
MOVEMENT }o--|| ADJUSTMENT : "source: ADJUSTMENT"
MOVEMENT }o--|| VENDOR_RMA : "source: RMA_IN/OUT"
MOVEMENT }o--|| STOCK_COUNT : "source: COUNT_ADJUST"
4.12.5 Reports: Movement History
| Report | Purpose | Key Data Fields |
|---|---|---|
| Product Movement Log | Full chronological history for one product at one location | Product, location, event type, qty change, running balance, source document, reference number, actor, timestamp |
| Location Movement Summary | Aggregate all movements at a location for a period | Location, period, event type, total events, total qty in, total qty out, net change |
| Shrinkage Analysis | Identify unexplained inventory loss | Location, period, expected balance (from ledger), actual balance (from count), unexplained variance, shrinkage %, shrinkage value (at weighted avg cost from Section 4.11) |
| Movement by Source | Volume of movements grouped by source document type | Source type (PO, Transfer, Sale, Adjustment, RMA), event count, total qty moved, period |
| Adjustment Audit Trail | All manual adjustments with reason codes | Adjustment reference, product, location, qty change, reason code, actor, timestamp, notes (see Section 4.7 for adjustment workflow) |
4.13 POS & Sales Integration
Scope: Real-time interaction between inventory management and point-of-sale operations. This section defines how inventory quantities are affected at each stage of a sales transaction, including cart operations, payment, voids, returns, and serial number capture.
4.13.1 Reserve on Add to Cart
When a cashier adds an item to a transaction, the system creates a soft reservation against the selling location’s available quantity. The reservation is tied to the terminal and transaction session.
Reservation behavior:
- Available qty at the selling location is decremented immediately in the UI and API layer.
- Other terminals at the same location see the reduced available quantity in real time.
- The reservation is temporary and tied to the active transaction session.
- If the item is removed from the cart, the reservation releases instantly.
- If the entire transaction is voided before payment, all reservations release instantly.
Data written on Add to Cart:
| Field | Value |
|---|---|
reservation_type | SOFT |
reservation_source | POS_CART |
terminal_id | Current terminal |
transaction_session_id | Active session |
product_variant_id | Selected variant |
location_id | Selling location |
qty_reserved | Quantity added |
created_at | Timestamp |
expires_at | NULL (cleared on payment or void) |
Cross-reference: See Section 1.1 (Core Sales Workflow) for the full item entry flow. Reservation logic integrates with the cart state machine defined there.
4.13.2 Commit on Payment
When payment completes successfully, the soft reservation converts to a permanent inventory decrement. A SALE movement is logged in the stock ledger (See Section 4.12).
Payment completion rules:
- On successful payment, the reservation record is deleted and replaced with a finalized
SALEmovement record. - The Weighted Average Cost (WAC) is captured at the moment of sale for COGS calculation.
- If payment fails (card decline, insufficient funds, terminal timeout), the reservation holds for 30 seconds then auto-releases.
- The 30-second hold prevents a race condition where another terminal could claim the last unit while the cashier retries payment.
- If the cashier retries payment within 30 seconds, the existing reservation is reused.
- After 30 seconds with no retry, the reservation releases and the item returns to available stock.
On payment success, the system writes:
| Record | Fields |
|---|---|
| Stock Movement | type: SALE, qty: -N, location_id, product_variant_id, reference_type: ORDER, reference_id: order_id, cost_at_time: WAC, created_by: staff_id |
| Inventory Level Update | available_qty -= N at selling location |
| Reservation Cleanup | Delete soft reservation record |
4.13.3 Reserve for Parked Transactions
Parked (suspended) sales maintain soft reservations throughout their lifecycle. The parked sale state machine (See Section 1.1.1) governs the reservation lifecycle.
Parked sale reservation rules:
| Rule | Value | Configurable |
|---|---|---|
| Reservation type | SOFT | No |
| Visibility to other terminals | Available with warning | No |
| Warning message format | “N units available, M reserved by Terminal X” | No |
| Maximum parked sales per terminal | 5 | Yes |
| Time-to-live (TTL) | 4 hours | Yes |
| On TTL expiry | Auto-release all reservations | No |
| On retrieval | Reservations transfer to active cart session | No |
Parked sale warning display example:
Product: Classic Fit Tee - Navy / M
Available: 2 units
Warning: 1 unit reserved by Terminal 3 (Parked Sale)
When a parked sale expires:
- All reservations release to available stock.
- A
RESERVATION_EXPIREDevent is logged with the parked sale ID. - The expired sale is archived with reason
TTL_EXCEEDED.
Cross-reference: See Section 1.1.1 (Parked Sale State Machine) for TTL configuration and the
parked_salesYAML config in Section 4.18.
4.13.4 Reserve for Hold-for-Pickup
Fully paid orders designated for customer pickup create hard reservations. These items are not visible to other terminals as available stock.
Hold-for-pickup reservation rules:
| Rule | Value | Configurable |
|---|---|---|
| Reservation type | HARD | No |
| Visibility to other terminals | Not visible as available | No |
| Inventory status | RESERVED | No |
| Default hold period | 7 days | Yes |
| Maximum hold extension | 30 days | Yes |
| Reminder before expiry | 2 days | Yes |
| On expiry | Auto-refund process triggers | Yes |
Hard reservation behavior:
- Available qty is decremented. The qty is moved to
reserved_qtyin the inventory level record. - Other terminals see only the
available_qty(excluding reserved). - The POS dashboard shows held orders with countdown timers.
- When the customer picks up, the reservation clears and the sale is marked
COMPLETED. - When the hold expires without pickup, the system initiates the auto-refund workflow and the inventory moves back to
AVAILABLEstatus.
Cross-reference: See Section 1.11 (Hold for Pickup) for the full pickup workflow and BOPIS integration. See
hold_for_pickupYAML config in Section 4.18.
4.13.5 Inventory Decrement on Sale
Each completed sale logs a SALE movement in the stock ledger. This is the authoritative record of inventory leaving the business through a customer transaction.
Movement record for a sale:
| Field | Value |
|---|---|
movement_type | SALE |
qty_change | Negative (e.g., -1) |
location_id | Selling location |
product_variant_id | Sold variant |
reference_type | ORDER |
reference_id | Order ID |
unit_cost_at_time | WAC at moment of sale |
created_by | Staff ID |
created_at | Transaction timestamp |
notes | NULL (auto-generated from sale) |
Multi-item transactions: Each line item in a transaction generates its own movement record. All movements for a single transaction share the same reference_id (order ID).
WAC capture: The system snapshots the current Weighted Average Cost at the moment of sale. This ensures COGS calculations remain accurate even if costs change later due to new receiving events.
Cross-reference: See Section 4.12 (Stock Ledger & Movement Log) for the complete movement type taxonomy.
4.13.6 Inventory Increment on Return
Customer returns automatically restore inventory at the return location. The default behavior places returned items into AVAILABLE status, but staff can override to DAMAGED if the item is not resalable.
Return-to-stock rules:
| Rule | Behavior |
|---|---|
| Default return status | AVAILABLE |
| Staff override options | DAMAGED, QUARANTINE |
| Movement type logged | RETURN |
| Location | Return location (may differ from sale location) |
| Cross-store returns | Allowed; inventory incremented at return location, not original sale location |
| WAC impact | No WAC recalculation on return (original cost basis preserved) |
| Serial-tracked items | Serial status reverts to IN_STOCK at return location |
| Lot-tracked items | Lot assignment restored; FIFO position maintained |
Movement record for a return:
| Field | Value |
|---|---|
movement_type | RETURN |
qty_change | Positive (e.g., +1) |
location_id | Return location |
product_variant_id | Returned variant |
reference_type | RETURN |
reference_id | Return transaction ID |
unit_cost_at_time | Original sale cost |
created_by | Staff ID |
notes | Reason code (e.g., DEFECTIVE, WRONG_SIZE, CHANGED_MIND) |
Cross-reference: See Section 1.4.1 (Void vs. Return Rules) for return eligibility logic and Section 4.7 (Vendor RMA) for items returned to vendor instead of restocked.
4.13.7 Serial Number Capture at POS
For products flagged as serial-tracked, the POS enforces serial number capture during both sale and return transactions.
Serial capture at sale:
- Cashier scans or adds a serial-tracked product to the cart.
- POS prompts: “Scan or enter serial number.”
- Cashier scans barcode or manually enters serial number.
- System validates:
- Serial number exists in the system (was recorded at receiving).
- Serial status is
IN_STOCKat the selling location. - Serial is not already linked to another active sale.
- On validation pass, serial is attached to the line item.
- On payment completion, serial status changes to
SOLDand is linked to the customer record.
Serial capture at return:
- Staff initiates return and scans the serial number.
- System looks up the serial and retrieves the original sale record.
- Serial status reverts to
IN_STOCKat the return location. - If staff marks item as damaged, serial status changes to
DAMAGED.
Serial validation errors:
| Error | Message | Action |
|---|---|---|
| Serial not found | “Serial number not recognized. Verify and retry.” | Block sale line |
| Serial already sold | “Serial already linked to Order #X. Investigate.” | Block sale line |
| Serial at wrong location | “Serial is at [Location]. Transfer required.” | Block sale line |
| Serial in quarantine | “Serial is quarantined. Manager override required.” | Require manager PIN |
Cross-reference: See Section 4.10 (Serial & Lot Tracking) for the full serial lifecycle and Section 1.10 (Serial Number Tracking) for POS-side serial workflows.
4.13.8 Sale Flow Sequence Diagram
sequenceDiagram
autonumber
participant U as Cashier
participant POS as POS Terminal
participant API as Inventory API
participant DB as Database
participant LED as Stock Ledger
Note over U, LED: Phase 1: Item Added to Cart
U->>POS: Scan Item (SKU-1001)
POS->>API: POST /inventory/reserve
API->>DB: Create soft reservation (terminal_id, session_id, product_variant_id, qty)
API->>DB: Decrement available_qty at location
DB-->>API: Reservation confirmed
API-->>POS: Reserved (available: 4 → 3)
POS-->>U: Item added to cart
Note over U, LED: Phase 2: Payment Processing
U->>POS: Tender Payment
POS->>API: POST /orders/finalize
API->>DB: Validate reservation still active
API->>DB: Delete soft reservation
API->>LED: INSERT movement (type: SALE, qty: -1, cost: WAC)
API->>DB: Update available_qty (permanent decrement)
DB-->>API: Sale committed
API-->>POS: Order finalized (Order #ORD-5678)
POS-->>U: Print receipt
Note over U, LED: Phase 3: Payment Failure Path
Note right of POS: If payment fails:
Note right of POS: Reservation holds 30 seconds
Note right of POS: Cashier retries → reuse reservation
Note right of POS: No retry within 30s → auto-release
4.13.9 Return Flow Sequence Diagram
sequenceDiagram
autonumber
participant U as Staff
participant POS as POS Terminal
participant API as Inventory API
participant DB as Database
participant LED as Stock Ledger
Note over U, LED: Phase 1: Return Initiated
U->>POS: Initiate Return
POS-->>U: "Scan receipt or enter order number"
U->>POS: Scan Receipt Barcode
POS->>API: GET /orders/{order_id}
API-->>POS: Order details + line items
Note over U, LED: Phase 2: Validate & Process
U->>POS: Select items to return
POS->>API: POST /returns/validate
API->>API: Check return policy (window, final sale, etc.)
API-->>POS: Return eligible
U->>POS: Confirm return + select condition
alt Item is resalable
POS->>API: POST /returns/process (status: AVAILABLE)
else Item is damaged
POS->>API: POST /returns/process (status: DAMAGED)
end
Note over U, LED: Phase 3: Inventory Updated
API->>DB: Increment available_qty (or damaged_qty) at return location
API->>LED: INSERT movement (type: RETURN, qty: +1, cost: original_cost)
API->>DB: Update order line status to RETURNED
DB-->>API: Return processed
API-->>POS: Return complete (Refund: $29.99)
POS-->>U: Process refund to original payment method
POS-->>U: Print return receipt
4.14 Online Order Fulfillment
Scope: How online orders placed through Shopify interact with physical store inventory, including store assignment, inventory reservation, bidirectional sync, and the pick-pack-ship workflow.
4.14.1 Reserve from Nearest Store
When a customer places an online order through Shopify, the system identifies the optimal fulfillment location based on proximity to the customer’s shipping address and stock availability.
Reservation flow:
- Shopify webhook fires
orders/createevent. - System receives the order and extracts the shipping address.
- Store assignment algorithm (Section 4.14.2) selects the fulfillment location.
- Inventory is hard reserved at the selected store (status:
RESERVED, source:ONLINE_ORDER). - The order appears on the selected store’s fulfillment queue.
- If no store has sufficient stock, the order is flagged for manual assignment.
Online order reservation record:
| Field | Value |
|---|---|
reservation_type | HARD |
reservation_source | ONLINE_ORDER |
shopify_order_id | Shopify order reference |
assigned_location_id | Selected store |
product_variant_id | Ordered variant |
qty_reserved | Order quantity |
status | PENDING_FULFILLMENT |
created_at | Order timestamp |
expires_at | NULL (does not expire; requires manual cancel) |
4.14.2 Store Assignment Algorithm
The store assignment algorithm determines which physical store fulfills an online order. The algorithm prioritizes proximity while ensuring stock availability.
Algorithm steps:
1. FILTER: Stores where available_qty >= ordered_qty for ALL line items
2. IF no single store has all items:
a. IF split_fulfillment_enabled = true:
- Find minimum set of stores to cover all items
- Prefer fewer splits (2 stores over 3)
- Within equal splits, prefer nearest stores
b. IF split_fulfillment_enabled = false:
- Flag order for manual assignment
- Notify HQ manager
3. CALCULATE: Distance from each qualifying store to customer shipping address
- Uses haversine formula on store lat/lng vs. shipping address lat/lng
4. SORT: By distance ascending
5. SELECT: Nearest qualifying store
6. RESERVE: Inventory at selected store
Store assignment decision matrix:
| Scenario | Single Store Available | Multiple Stores Available | No Store Available |
|---|---|---|---|
| Full order at one store | Assign nearest | Assign nearest with full stock | Flag for manual |
| Partial at multiple stores | Flag for manual (split disabled) | Split across nearest stores (split enabled) | Flag for manual |
| HQ warehouse only | Assign HQ | Prefer retail store over HQ | Backorder or cancel |
flowchart TD
A[Online Order Received] --> B{Any store has\nall items?}
B -->|Yes| C[Filter stores with full stock]
C --> D[Calculate distance to each store]
D --> E[Select nearest store]
E --> F[Reserve inventory at store]
F --> G[Order appears on\nstore fulfillment queue]
B -->|No| H{Split fulfillment\nenabled?}
H -->|Yes| I[Find minimum store\nset covering all items]
I --> J[Create split shipments]
J --> K[Reserve at each store]
K --> G
H -->|No| L{HQ has stock?}
L -->|Yes| M[Assign to HQ warehouse]
M --> F
L -->|No| N[Flag for manual\nassignment]
N --> O[Notify HQ manager]
4.14.3 Inventory Sync with Shopify
MOVED TO MODULE 6: Shopify inventory sync triggers, architecture, and reconciliation details have been consolidated into Module 6, Section 6.3.14 (Inventory Sync Triggers) and Section 6.7 (Cross-Platform Inventory Sync Rules).
See: Module 6, Section 6.3.14 for Shopify-specific inventory sync triggers and Section 6.7 for cross-platform inventory sync architecture including safety buffers and oversell prevention.
4.14.4 Pick-Pack-Ship Workflow
Once an online order is assigned to a store, the store staff fulfills it through a structured pick-pack-ship process.
Workflow stages:
| Stage | Action | System Effect |
|---|---|---|
| 1. Order Received | Order appears on store’s fulfillment queue | Status: PENDING_FULFILLMENT. Inventory reserved. |
| 2. Pick | Staff locates and scans each item | Status: PICKING. System validates each scanned item against the order. |
| 3. Pack | Staff packages items for shipping | Status: PACKING. Staff selects box size and records weight. |
| 4. Ship | Staff enters carrier and tracking number | Status: SHIPPED. Inventory decremented (SALE movement logged). Shopify order marked fulfilled with tracking. |
| 5. Deliver | Carrier delivers to customer | Status: DELIVERED. Updated via carrier webhook or manual confirmation. |
Pick validation rules:
- Each item must be scanned individually.
- If scanned item does not match order line, system rejects with “Item not on this order.”
- If item is serial-tracked, serial number must be captured during pick.
- If item is lot-tracked, system auto-selects FIFO lot and records lot number.
- Staff can flag a “short pick” if an item is not found. This triggers a recount at the store and potential reassignment to another store.
sequenceDiagram
autonumber
participant SH as Shopify
participant API as POS API
participant DB as Database
participant ST as Store Staff
participant CR as Carrier
Note over SH, CR: Phase 1: Order Assignment
SH->>API: Webhook: orders/create
API->>API: Run store assignment algorithm
API->>DB: Reserve inventory at selected store
API->>DB: Create fulfillment record (PENDING_FULFILLMENT)
API-->>ST: New order on fulfillment queue
Note over SH, CR: Phase 2: Pick & Pack
ST->>API: Start picking (fulfillment_id)
API->>DB: Status: PICKING
loop Each order line
ST->>API: Scan item barcode
API->>DB: Validate item matches order line
API-->>ST: Item confirmed
end
ST->>API: Picking complete
API->>DB: Status: PACKING
ST->>API: Record package details (weight, dimensions)
API->>DB: Status: READY_TO_SHIP
Note over SH, CR: Phase 3: Ship & Track
ST->>API: Enter carrier + tracking number
API->>DB: Status: SHIPPED
API->>DB: Log SALE movement for each line item
API->>DB: Decrement inventory (release reservation, permanent decrement)
API->>SH: POST fulfillment with tracking number
SH-->>API: Fulfillment confirmed
CR->>API: Webhook: delivered
API->>DB: Status: DELIVERED
4.15 Offline Inventory Operations
Scope: How inventory operations function when a store loses network connectivity to the central server. This section defines which operations continue locally, which are blocked, and how conflicts are resolved upon reconnection.
4.15.1 Queue All Changes
During offline mode, the POS client maintains a local inventory cache and queues all changes for synchronization when connectivity restores. The local cache is updated immediately so staff can continue working.
Queued operations (allowed offline):
| Operation | Local Cache Effect | Queue Entry |
|---|---|---|
| Sale decrement | available_qty -= N in local cache | { type: SALE, product_variant_id, location_id, qty: -N, order_ref, timestamp } |
| Return increment | available_qty += N in local cache | { type: RETURN, product_variant_id, location_id, qty: +N, return_ref, timestamp } |
| Inventory adjustment | available_qty += delta in local cache | { type: ADJUSTMENT, product_variant_id, location_id, qty: +/-N, reason_code, timestamp } |
| Stock count entry | Stored in local count session | { type: COUNT, count_session_id, product_variant_id, counted_qty, timestamp } |
| Parked sale create | Reservation in local cache | { type: PARK, session_id, items: [...], terminal_id, timestamp } |
| Parked sale retrieve | Reservation transferred in local cache | { type: UNPARK, session_id, terminal_id, timestamp } |
Local cache structure:
local_inventory_cache = {
product_variant_id: {
available_qty: number, // Last synced value +/- local changes
reserved_qty: number, // Local reservations
last_synced_at: timestamp, // When cache was last updated from server
pending_changes: [ // Ordered list of unsynced changes
{ type, qty, ref, timestamp }
]
}
}
Queue limits:
- Maximum queued inventory changes: 500 entries (configurable).
- Maximum offline duration before warning: 4 hours.
- If queue reaches 90% capacity, POS displays warning: “Offline queue nearly full. Reconnect soon.”
- If queue reaches 100%, POS blocks further inventory-modifying operations.
4.15.2 Blocked Operations
The following operations require server connectivity and are blocked during offline mode. The POS displays a clear message explaining why the operation is unavailable.
| Blocked Operation | Reason | User Message |
|---|---|---|
| Multi-store inventory lookup | Requires real-time data from all locations | “Multi-store lookup unavailable offline. Check local stock only.” |
| Cross-store transfer request | Requires server to coordinate with other stores | “Transfers unavailable offline. Queue request when online.” |
| Online order fulfillment | Requires Shopify sync and server coordination | “Online fulfillment unavailable offline.” |
| Shopify inventory sync | Requires Shopify API connectivity | Syncs automatically on reconnect. |
| PO submission to vendor | Requires email delivery and server logging | “PO submission unavailable offline. Save as draft.” |
| PO receiving | Requires WAC recalculation and server-side validation | “Receiving unavailable offline. Wait for connectivity.” |
| Gift card activation/reload | Requires server validation of card status | “Gift card operations unavailable offline.” |
| Customer account creation | Requires server-side deduplication | “Customer creation unavailable offline.” |
| RMA creation | Requires server-side RMA number generation | “RMA creation unavailable offline.” |
Cross-reference: See Section 1.16 (Offline Operations) for the complete offline mode state machine and the
offline_modeYAML config in the BRD Section 4 Business Rules.
4.15.3 Conflict Resolution on Reconnect
When connectivity restores, the system uploads all queued changes and reconciles them with changes that occurred at other locations during the offline period.
Sync process (step by step):
1. POS detects network restored → Status: SYNCING
2. Upload queued changes in strict chronological order (oldest first)
3. For each queued change:
a. Server calculates expected qty = (last_synced_qty + all server-side changes since sync)
b. Server applies the offline change
c. IF resulting qty < 0 AND allow_negative_inventory = false:
- Flag as CONFLICT
- Record: { product_variant_id, expected_qty, offline_change, resulting_qty, conflict_type: NEGATIVE_INVENTORY }
d. IF resulting qty >= 0:
- Apply change normally
- Log movement with offline_synced: true flag
4. After all changes processed:
a. If conflicts exist → Status: CONFLICT_REVIEW
b. If no conflicts → Status: ONLINE
5. Manager reviews and resolves each conflict
Conflict resolution rules:
| Conflict Type | Resolution Strategy | Manager Action |
|---|---|---|
| Negative inventory (same item sold at two stores) | Last-write-wins for quantities | Review and approve negative balance, or adjust |
| Count submitted offline while count also submitted at server | Server count wins; offline count flagged for re-review | Accept server count or request recount |
| Adjustment conflicts | Both adjustments applied; resulting qty may need review | Verify physical count matches |
| Parked sale item no longer available | Auto-release parked sale reservation | Notify customer, offer alternatives |
Conflict review dashboard fields:
| Field | Description |
|---|---|
conflict_id | Unique conflict identifier |
product_variant_id | Affected product |
location_id | Affected store |
offline_terminal_id | Terminal that made the offline change |
offline_change | The queued change (type, qty, timestamp) |
server_qty_at_sync | Server’s qty when sync occurred |
resulting_qty | Calculated result after applying offline change |
conflict_type | NEGATIVE_INVENTORY, COUNT_CONFLICT, ADJUSTMENT_CONFLICT |
resolution | PENDING, ACCEPTED, ADJUSTED, REVERSED |
resolved_by | Manager who resolved |
resolved_at | Resolution timestamp |
sequenceDiagram
autonumber
participant POS as POS Terminal
participant Q as Local Queue
participant API as Server API
participant DB as Database
participant MGR as Manager
Note over POS, MGR: Phase 1: Offline Operations
POS->>POS: Detect network lost → OFFLINE MODE
POS->>Q: Queue sale: SKU-1001, qty: -1, 10:15 AM
POS->>Q: Queue sale: SKU-1001, qty: -2, 10:32 AM
POS->>Q: Queue adjustment: SKU-2005, qty: -3, 11:00 AM
Note right of Q: Meanwhile, Store B sells 2x SKU-1001 on server
Note over POS, MGR: Phase 2: Reconnect & Sync
POS->>POS: Detect network restored → SYNCING
POS->>API: Upload queue (3 entries, chronological)
API->>DB: Apply sale SKU-1001 qty: -1 (10:15 AM)
DB-->>API: OK (server qty was 5, now 4)
API->>DB: Apply sale SKU-1001 qty: -2 (10:32 AM)
DB-->>API: OK (server qty was 4, now 2)
API->>DB: Apply adjustment SKU-2005 qty: -3 (11:00 AM)
DB-->>API: CONFLICT: resulting qty = -1
API-->>POS: Sync result: 2 OK, 1 conflict
Note over POS, MGR: Phase 3: Conflict Resolution
POS->>POS: Status → CONFLICT_REVIEW
POS-->>MGR: Alert: 1 inventory conflict requires review
MGR->>API: Review conflict: SKU-2005, expected qty 2, adjustment -3 = -1
MGR->>API: Resolution: Accept negative, schedule recount
API->>DB: Apply adjustment, flag for investigation
API-->>POS: All conflicts resolved → ONLINE
4.16 Alerts, Notifications & Email Templates
Scope: Proactive inventory alerts that notify staff and managers of conditions requiring attention, plus automated email templates for key inventory events.
4.16.1 Alert Types
The system supports five inventory alert types, each with configurable thresholds, severity levels, and delivery channels.
| Alert | Trigger | Severity | Delivery | Configurable Parameters |
|---|---|---|---|---|
| Low Stock | Qty falls below reorder point for product at location | WARNING | Dashboard + daily email digest | Reorder point per product per location; digest send time |
| Overstock | Days of supply exceeds threshold | INFO | Dashboard only | Days of supply threshold (default: 90 days) |
| Shrinkage Threshold | Count variance exceeds % threshold of expected qty | CRITICAL | Dashboard + immediate email | Variance % threshold (default: 5%) |
| Aging Inventory | No sales for product at location in X days | WARNING | Dashboard + weekly email digest | Days threshold (default: 90 days); weekly digest day (default: Monday) |
| PO Overdue | PO not received within vendor lead time + buffer days | WARNING | Dashboard + email to buyer | Buffer days beyond lead time (default: 3 days) |
Alert trigger logic (detailed):
Low Stock:
FOR each (product_variant_id, location_id):
IF available_qty <= reorder_point
AND no active OPEN alert exists for this product/location combo
THEN create LOW_STOCK alert
Overstock:
FOR each (product_variant_id, location_id):
days_of_supply = available_qty / avg_daily_sales_velocity_90d
IF days_of_supply > overstock_threshold_days
AND avg_daily_sales_velocity_90d > 0 -- exclude dead stock (handled separately)
THEN create OVERSTOCK alert
Shrinkage Threshold:
ON count finalization:
FOR each counted product:
variance_pct = ABS(counted_qty - expected_qty) / expected_qty * 100
IF variance_pct > shrinkage_threshold_pct
THEN create SHRINKAGE alert (CRITICAL)
Aging Inventory:
WEEKLY job:
FOR each (product_variant_id, location_id):
last_sale_date = MAX(sale_date) for this product at this location
days_since_sale = TODAY - last_sale_date
IF days_since_sale > aging_threshold_days
AND available_qty > 0
THEN create AGING_INVENTORY alert
PO Overdue:
DAILY job:
FOR each PO in status OPEN or PARTIAL:
expected_date = po.created_at + vendor.lead_time_days + buffer_days
IF TODAY > expected_date
AND no active PO_OVERDUE alert exists for this PO
THEN create PO_OVERDUE alert
4.16.2 Alert Data Model
| Field | Type | Description |
|---|---|---|
alert_id | UUID | Unique alert identifier |
tenant_id | UUID | Tenant scope |
alert_type | ENUM | LOW_STOCK, OVERSTOCK, SHRINKAGE, AGING_INVENTORY, PO_OVERDUE |
severity | ENUM | INFO, WARNING, CRITICAL |
product_variant_id | UUID | Affected product (nullable for PO alerts) |
location_id | UUID | Affected location |
reference_type | VARCHAR | INVENTORY_LEVEL, COUNT_SESSION, PURCHASE_ORDER |
reference_id | UUID | ID of the triggering entity |
message | TEXT | Human-readable alert description |
data_snapshot | JSONB | Key metrics at alert time (e.g., { "available_qty": 2, "reorder_point": 5 }) |
triggered_at | TIMESTAMP | When alert was created |
acknowledged_by | UUID | Staff who acknowledged (nullable) |
acknowledged_at | TIMESTAMP | When acknowledged (nullable) |
resolved_at | TIMESTAMP | When condition cleared (nullable) |
auto_resolved | BOOLEAN | true if resolved by system (e.g., stock replenished) |
4.16.3 Alert Lifecycle
stateDiagram-v2
[*] --> TRIGGERED: Condition detected
TRIGGERED --> ACKNOWLEDGED: Staff views/clicks alert
ACKNOWLEDGED --> RESOLVED: Condition cleared (manual or auto)
TRIGGERED --> RESOLVED: Condition auto-clears (e.g., stock received)
RESOLVED --> [*]
note right of TRIGGERED
Appears on dashboard
May send email/notification
end note
note right of ACKNOWLEDGED
Staff is aware
Working on resolution
end note
note right of RESOLVED
Condition no longer active
Retained for history/reporting
end note
Auto-resolution rules:
| Alert Type | Auto-Resolves When |
|---|---|
| Low Stock | available_qty > reorder_point (stock received or transferred in) |
| Overstock | days_of_supply <= overstock_threshold (stock sold or transferred out) |
| Shrinkage | Never auto-resolves; requires manual acknowledgment |
| Aging Inventory | Sale occurs for the product at the location |
| PO Overdue | PO status changes to RECEIVED or COMPLETED |
4.16.4 Email Templates
Four email templates cover the primary inventory communication needs. Each template uses dynamic field substitution.
Template 1: TMPL_INV_PO_VENDOR
| Property | Value |
|---|---|
| Template ID | TMPL_INV_PO_VENDOR |
| Trigger | PO status changes to OPEN (submitted to vendor) |
| Recipient | Vendor email address (from vendor record) |
| Subject | Purchase Order {po_number} from {tenant_name} |
| Dynamic Fields | vendor_name, ship_to_address (store or HQ), po_number, po_date, expected_delivery_date, line_items_table (SKU, description, qty, unit cost, line total), po_total, special_instructions, buyer_name, buyer_email, buyer_phone |
Template 2: TMPL_INV_TRANSFER_ALERT
| Property | Value |
|---|---|
| Template ID | TMPL_INV_TRANSFER_ALERT |
| Trigger | Transfer status changes to SHIPPED |
| Recipient | Destination store manager email |
| Subject | Incoming Transfer {transfer_number} from {source_store_name} |
| Dynamic Fields | source_store_name, destination_store_name, transfer_number, shipped_date, expected_arrival_date, manifest_table (SKU, description, qty shipped), total_items, tracking_number, carrier_name, shipper_name |
Template 3: TMPL_INV_LOW_STOCK
| Property | Value |
|---|---|
| Template ID | TMPL_INV_LOW_STOCK |
| Trigger | Daily job (configurable time, default: 7:00 AM) |
| Recipient | Store manager and/or HQ inventory manager (configurable) |
| Subject | Low Stock Alert: {count} items below reorder point at {location_name} |
| Dynamic Fields | location_name, report_date, count (number of items), items_table (SKU, product name, current qty, reorder point, reorder qty, primary vendor, vendor lead time), total_reorder_value |
Template 4: TMPL_INV_COUNT_REMINDER
| Property | Value |
|---|---|
| Template ID | TMPL_INV_COUNT_REMINDER |
| Trigger | Scheduled count is N days away (configurable, default: 2 days) |
| Recipient | Store manager |
| Subject | Inventory Count Reminder: {count_type} scheduled for {scheduled_date} |
| Dynamic Fields | count_type (Full, Cycle, Spot Check), scheduled_date, scheduled_time, location_name, scope_description (e.g., “Category: Tops” or “Full Store”), assigned_staff_names, estimated_duration, special_instructions |
4.17 Inventory Dashboard & Reports
Scope: A dedicated inventory dashboard providing at-a-glance KPIs and a comprehensive reporting suite consolidating all reports referenced across Sections 4.3 through 4.16.
4.17.1 Dashboard KPIs
The inventory dashboard displays eight primary KPI cards in a responsive grid layout. Each card shows the current value, trend indicator, and drill-down link.
| KPI Card | Metric | Trend | Drill-Down |
|---|---|---|---|
| Total Inventory Value (WAC) | Sum of (available_qty x WAC) across all locations | 30-day trend arrow (up/down/flat) + % change | Inventory Valuation report |
| Low Stock Items | Count of products where available_qty <= reorder_point | Delta from prior week | Low Stock Alert report |
| Pending PO Count | Count of POs in OPEN or PARTIAL status | Total pending PO value in parentheses | Open PO Report |
| Open Transfers | Count of transfers in REQUESTED, APPROVED, PICKING, or SHIPPED status | Total in-transit items count | Open Transfer Report |
| Upcoming Counts | Count of scheduled counts in next 7 days | Next count date and type | Count schedule calendar |
| Shrinkage % | (Total variance value / Total inventory value) x 100 for last 30 days | vs. prior 30-day period | Shrinkage Analysis report |
| Dead Stock Count | Count of products with zero sales velocity in last 90 days | Delta from prior month | Dead Stock Report |
| Avg Days of Supply | Average days_of_supply across all active products by category | Top 3 categories with lowest supply | Days of Supply report |
Dashboard filters:
- Location: All Stores, specific store, or HQ
- Category: All, or specific category
- Date range: Applies to trend calculations
- Brand: All, or specific brand
4.17.2 Master Report Suite
The following table consolidates all inventory reports from across the module. Each report includes its source section, purpose, key data fields, grouping keys, and export formats.
| # | Report Name | Source Section | Purpose | Key Data Fields | Grouping Keys | Export |
|---|---|---|---|---|---|---|
| 1 | Inventory Snapshot | 4.3 | Current QoH and total value at a point in time | SKU, product name, variant, location, available qty, reserved qty, total qty, WAC, extended value | Location, Category, Brand, Vendor | CSV, PDF |
| 2 | Low Stock Alert | 4.5 | Items below reorder point requiring restock action | SKU, product name, location, available qty, reorder point, reorder qty, primary vendor, lead time | Location, Vendor, Category | CSV, PDF, Dashboard |
| 3 | Stock Movement Log | 4.12 | Full audit trail of all inventory ins and outs | Movement ID, timestamp, type, SKU, product name, qty change, location, reference type, reference ID, staff, notes | Movement Type, Location, Date Range, Staff | CSV |
| 4 | Inventory Valuation (WAC) | 4.11 | Total inventory value using Weighted Average Cost | SKU, product name, location, qty, WAC per unit, extended value, % of total value | Location, Category, Brand | CSV, PDF |
| 5 | Shrinkage Analysis | 4.6 | Expected vs. actual quantities from count sessions | Count session, date, location, SKU, expected qty, counted qty, variance, variance %, variance value | Location, Date Range, Category | CSV, PDF |
| 6 | Vendor Performance Scorecard | 4.4 | Vendor reliability metrics across POs | Vendor name, total POs, on-time %, avg lead time, fill rate %, defect rate %, cost variance %, total spend | Vendor, Date Range | CSV, PDF |
| 7 | Inventory Turnover | 4.5 | Stock efficiency metrics | SKU, product name, category, avg inventory, COGS, turnover rate, days of supply, sell-through % | Category, Brand, Location | CSV, PDF |
| 8 | ABC Classification | 4.5 | Pareto analysis of products by revenue contribution | SKU, product name, revenue (period), cumulative revenue %, classification (A/B/C), qty sold, avg margin | Category, Brand, Location | CSV, PDF |
| 9 | Aging Analysis | 4.5 | Inventory age distribution by time buckets | SKU, product name, location, qty, receive date, age (days), age bucket (0-30, 31-60, 61-90, 90+), value | Location, Category, Age Bucket | CSV, PDF |
| 10 | Dead Stock Report | 4.5 | Items with zero sales velocity over threshold period | SKU, product name, location, available qty, value, last sale date, days since last sale, receive date | Location, Category, Vendor | CSV, PDF |
| 11 | Open PO Report | 4.4 | Purchase orders pending receipt | PO number, vendor, status, created date, expected date, total lines, received lines, total value, outstanding value | Vendor, Status, Location | CSV, PDF |
| 12 | PO Receiving Report | 4.4 | Details of received PO line items | PO number, receive date, SKU, ordered qty, received qty, variance, unit cost, cost variance, receiver staff | PO Number, Vendor, Date Range | CSV, PDF |
| 13 | Vendor Lead Time Report | 4.4 | Actual vs. expected delivery times by vendor | Vendor name, PO number, ordered date, expected date, received date, actual lead time, variance (days) | Vendor, Date Range | CSV, PDF |
| 14 | PO Variance Report | 4.4 | Discrepancies between ordered and received | PO number, SKU, ordered qty, received qty, qty variance, ordered cost, actual cost, cost variance | PO Number, Vendor | CSV, PDF |
| 15 | Cost Analysis Report | 4.11 | WAC trends and cost changes over time | SKU, product name, WAC (current), WAC (30d ago), WAC (90d ago), cost change %, last PO cost, vendor | Category, Vendor, Date Range | CSV, PDF |
| 16 | Reorder Alerts | 4.5 | Products approaching or at reorder point | SKU, product name, location, available qty, reorder point, days of supply, velocity, recommended qty, primary vendor | Location, Vendor, Category | CSV, PDF, Dashboard |
| 17 | Auto-PO Performance | 4.5 | Effectiveness of automatic PO generation | Month, auto-PO count, manual PO count, auto-PO accuracy %, stock-out events, avg days to stock-out | Month, Location | CSV, PDF |
| 18 | Velocity Trends | 4.5 | Sales velocity over time per product | SKU, product name, velocity (7d), velocity (30d), velocity (90d), trend direction, seasonality index | Category, Brand, Location | CSV, PDF |
| 19 | Days of Supply | 4.5 | How long current stock will last at current velocity | SKU, product name, location, available qty, avg daily velocity, days of supply, classification | Location, Category, Risk Level | CSV, PDF |
| 20 | Open RMA Report | 4.7 | Vendor returns in progress | RMA number, vendor, status, created date, total lines, total units, total value, expected credit | Vendor, Status | CSV, PDF |
| 21 | Vendor Return Rate | 4.7 | Defect and return rates by vendor | Vendor name, total received units, returned units, return rate %, top return reasons, credit recovered | Vendor, Date Range | CSV, PDF |
| 22 | RMA Aging | 4.7 | RMA age analysis for follow-up | RMA number, vendor, status, created date, age (days), last action date, total value, expected credit | Vendor, Status, Age Bucket | CSV, PDF |
| 23 | RMA Credit Reconciliation | 4.7 | Expected vs. received vendor credits | RMA number, vendor, expected credit, received credit, variance, credit date, reconciliation status | Vendor, Date Range, Status | CSV, PDF |
| 24 | Open Transfer Report | 4.8 | Transfers in progress between stores | Transfer number, source, destination, status, created date, shipped date, total items, expected arrival | Status, Source, Destination | CSV, PDF |
| 25 | Transfer Variance Report | 4.8 | Discrepancies between shipped and received quantities | Transfer number, SKU, shipped qty, received qty, variance, variance reason, source, destination | Transfer Number, Location | CSV, PDF |
| 26 | Transfer Volume Report | 4.8 | Transfer activity metrics over time | Period, total transfers, total units moved, avg transit days, top source locations, top destination locations | Date Range, Location Pair | CSV, PDF |
| 27 | Rebalancing Suggestions | 4.8 | System-generated transfer recommendations | SKU, product name, source location, source qty, source days of supply, destination location, destination qty, destination days of supply, suggested transfer qty | Category, Priority | CSV, PDF |
| 28 | Serial Number Lookup | 4.10 | Complete history of a serial-tracked unit | Serial number, SKU, product name, current status, current location, receive date, PO number, sale date, order ID, customer, return history | Status, Location | CSV, PDF |
| 29 | Lot Inventory Report | 4.10 | Current stock by lot number | Lot number, SKU, product name, location, qty available, receive date, expiry date (if applicable), age (days), PO number | Location, SKU, Expiry Status | CSV, PDF |
| 30 | Lot Trace (Recall) | 4.10 | Full traceability for recall management | Lot number, SKU, received qty, receive date, PO number, vendor, sold qty, remaining qty, customer list (with order IDs), locations distributed to | Lot Number | CSV, PDF |
| 31 | Landed Cost Analysis | 4.11 | Cost breakdown including freight, duty, and handling | PO number, SKU, base unit cost, freight allocation, duty allocation, handling allocation, total landed cost, allocation method | Vendor, PO Number | CSV, PDF |
| 32 | Margin Analysis | 4.11 | Gross margin by product and category | SKU, product name, category, sell price, WAC (landed), gross margin $, gross margin %, units sold, total margin | Category, Brand, Location | CSV, PDF |
| 33 | Cost Trend Report | 4.11 | Historical cost changes per product | SKU, product name, vendor, cost (current), cost (3m ago), cost (6m ago), cost (12m ago), trend, % change (YoY) | Vendor, Category, Trend Direction | CSV, PDF |
4.17.3 Report Access Control
| Role | View | Export | Schedule | Create Custom |
|---|---|---|---|---|
| Admin | All reports | All formats | Yes | Yes |
| HQ Manager | All reports | All formats | Yes | No |
| Store Manager | Store-scoped reports | CSV, PDF | Yes | No |
| Buyer | PO and vendor reports | CSV, PDF | Yes | No |
| Staff | Inventory Snapshot, Serial Lookup only | CSV only | No | No |
4.18 Inventory Business Rules — YAML Configuration
Cross-Reference: All inventory business rules configuration has been consolidated into Module 5: Setup & Configuration, Section 5.19.4 (Inventory Configuration). See Module 5, Section 5.19 for the complete YAML configuration reference covering all modules.
4.19 Inventory User Stories & Acceptance Criteria
Scope: All user stories organized by epic and Gherkin acceptance criteria for the Inventory Module. Epics 4.A through 4.F are moved from BRD Section 3.23 (renumbered). Epics 4.G through 4.P are new.
Epic 4.A: Vendor RMA
(Moved from Epic 3.I, renumbered)
Story 4.A.1: Create RMA
- As a Store Manager, I want to create a Return Merchandise Authorization with line items (product, qty, reason) so that I can return defective or overstock items to the vendor.
- Constraint: RMA numbers auto-increment per tenant using format
RMA-{YEAR}-{SEQUENCE}.
Story 4.A.2: RMA Workflow
- As a Buyer, I want an RMA to follow Draft > Submitted > Vendor Approved > Shipped Back > Credit/Replacement Received > Closed lifecycle so that vendor returns are tracked through every stage.
- Constraint: Each status transition is logged with timestamp and staff ID.
Story 4.A.3: Credit/Replacement
- As a Buyer, I want to record vendor credits against the RMA and track replacement shipments so that I can reconcile vendor credits with expected amounts.
- Constraint: When vendor sends replacement, a linked PO is created for receiving. Credit variance > 5% triggers alert.
Epic 4.B: Reorder Management
(Moved from Epic 3.J, renumbered)
Story 4.B.1: Dynamic Reorder Points
- As an Inventory Manager, I want the system to calculate reorder points from 90-day sales velocity, lead time, and safety stock so that reorder points stay current without manual maintenance.
- Constraint: Recalculated weekly via background job. Safety stock uses 1.65 sigma (95% service level).
Story 4.B.2: Auto-PO Generation
- As a Buyer, I want the system to auto-create draft POs when stock drops below reorder point so that I can review and submit without manually building each PO.
- Constraint: Staff reviews before submission. POs consolidated by vendor. Minimum PO value enforced.
Story 4.B.3: Seasonal Adjustment
- As an Inventory Manager, I want reorder velocity to adjust based on historical same-period data so that seasonal demand patterns are accounted for in reorder calculations.
- Constraint: Uses trailing 3-year same-month average when available. New products without history use category-level seasonality.
Epic 4.C: Inventory Control
(Moved from Epic 3.K, renumbered)
Story 4.C.1: Inventory Status
- As a Store Manager, I want each inventory unit to have a status (Available, Quarantine, Damaged, In-Transit, Reserved, On-Hold) so that only sellable stock is available for sale or transfer.
- Constraint: Only AVAILABLE status allows sale. Status transitions are logged to the movement history.
Story 4.C.2: Stock Counting
- As a Store Manager, I want five counting methods (full physical count, cycle count, scanner-assisted count, monthly spot check, on-demand count) so that I can choose the right method for each situation.
- Constraint: Workflow is Count > Review Variances > Approve Adjustments. High-variance items require manager approval.
Story 4.C.3: Adjustments
- As a Staff Member, I want to submit manual inventory adjustments with a reason code so that discrepancies can be corrected and tracked.
- Constraint: Reason codes required (DAMAGED, THEFT, COUNT_CORRECTION, VENDOR_RETURN, OTHER). Adjustments above configurable threshold require manager approval.
Story 4.C.4: Unified Receiving
- As a Warehouse Clerk, I want a single receiving workflow that handles all receiving types (PO receive, transfer receive, customer return, RMA replacement) so that the process is consistent regardless of source.
- Constraint: Barcode scanner verification. Variance tracking against expected quantities.
Story 4.C.5: Bulk Operations
- As a Buyer, I want to import products via CSV, export catalog data to CSV/XLSX, and make bulk changes to price/cost/status/category so that large-scale updates are efficient.
- Constraint: Bulk changes above configurable thresholds require approval workflow integration.
Epic 4.D: Inter-Store Transfers
(Moved from Epic 3.L, renumbered)
Story 4.D.1: Transfer Workflow
- As a Store Manager, I want inter-store transfers to follow Request > Approve > Pick > Ship > Receive > Complete lifecycle with variance tracking at each stage so that transfer accuracy is maintained.
- Constraint: Shipped qty vs. received qty variance logged. Destination must scan-confirm received items.
Story 4.D.2: Auto-Rebalancing
- As an HQ Manager, I want the system to analyze velocity vs. stock across locations and suggest transfers to equalize days of supply so that no store is overstocked while another is understocked.
- Constraint: Staff reviews and approves suggested transfers. Minimum imbalance threshold is configurable (default: 14 days difference).
Epic 4.E: Serial & Lot Tracking
(Moved from Epic 3.M, renumbered)
Story 4.E.1: Serial Tracking
- As a Cashier, I want serial-tracked products to require serial number entry at both receive and sale so that each unit is individually tracked for warranty and after-sale support.
- Constraint: Serial number linked to customer on sale. Serial validated as IN_STOCK at selling location before sale proceeds.
Story 4.E.2: Lot Tracking
- As a Warehouse Clerk, I want lot numbers assigned at receiving with FIFO enforcement on sale so that lot-tracked products are sold in order and full traceability is available for recall management.
- Constraint: FIFO enforced automatically. Lot trace report shows all customers who received items from a specific lot.
Epic 4.F: Landed Cost & Costing
(Moved from Epic 3.N, renumbered)
Story 4.F.1: Landed Cost
- As a Buyer, I want PO receiving to include cost allocation for freight, duties, customs, and handling so that the true per-unit cost is known for accurate margin calculations.
- Constraint: Three allocation methods supported (By Value, By Quantity, Manual). Landed cost stored as the true cost basis.
Story 4.F.2: Weighted Average Cost
- As an Inventory Manager, I want the system to maintain weighted average cost per product per location, recalculated on each receive, so that COGS calculations and margin reporting are accurate.
- Constraint: WAC formula:
New WAC = ((Existing Qty x Existing WAC) + (Received Qty x Received Cost)) / (Existing Qty + Received Qty). WAC is used for all COGS and margin calculations.
Epic 4.G: Receiving & Inspection
(New)
Story 4.G.1: Open Receive with Discrepancy Handling
- As a Warehouse Clerk, I want to receive a quantity different from the PO ordered quantity so that I can handle partial shipments, over-shipments, and damaged goods.
- Constraint: Uses the “triple approach” – (1) accept what arrived, (2) quarantine damaged items, (3) auto-create RMA for damages. PO status updates to PARTIAL if not all lines received.
Story 4.G.2: Non-PO Receiving
- As a Store Manager, I want to receive stock without a linked PO (e.g., consignment, found stock, vendor replacement) so that all inventory entering the store is tracked regardless of source.
- Constraint: Reason code required. Creates a standalone receiving record with movement log entry.
Story 4.G.3: Over-Shipment Threshold
- As a Buyer, I want the system to enforce a configurable over-receive threshold so that warehouses cannot accept significantly more than what was ordered without authorization.
- Constraint: Default threshold is 10% above PO line qty. Over-receive above threshold requires manager approval. Over-receive below threshold is accepted and logged.
Epic 4.H: POS Integration
(New)
Story 4.H.1: Reserve on Cart Add
- As a Cashier, I want inventory to be soft-reserved when I add an item to the cart so that another terminal at the same store does not sell the last unit before I complete payment.
- Constraint: Soft reservation decrements available qty immediately. Removing item from cart releases reservation instantly. Other terminals see reduced qty.
Story 4.H.2: Commit on Payment
- As a Cashier, I want the inventory reservation to convert to a permanent decrement when payment completes so that the stock ledger accurately reflects completed sales.
- Constraint: SALE movement logged. WAC captured at time of sale. If payment fails, reservation holds for 30 seconds then auto-releases.
Story 4.H.3: Return to Stock
- As a Store Staff Member, I want customer returns to automatically restore inventory to AVAILABLE status at the return location so that returned items are immediately available for resale.
- Constraint: Default status is AVAILABLE. Staff can override to DAMAGED if item is not resalable. RETURN movement logged. WAC is not recalculated (original cost preserved).
Epic 4.I: Online Fulfillment
(New)
Story 4.I.1: Nearest Store Reservation
- As an Online Operations Manager, I want online orders to automatically reserve inventory at the nearest store with stock so that orders ship quickly from the closest location.
- Constraint: Store assignment uses distance calculation from store to customer shipping address. If no single store has full stock and split fulfillment is disabled, order flags for manual assignment.
Story 4.I.2: Shopify Inventory Sync
- As an Online Operations Manager, I want inventory quantities to always sync bidirectionally between POS and Shopify so that online customers see accurate availability.
- Constraint: Webhook-driven sync (< 5 second target). Reconciliation every 15 minutes. POS is source of truth. Sync operates independently of catalog sync mode.
Story 4.I.3: Pick-Pack-Ship Workflow
- As a Store Staff Member, I want online orders assigned to my store to appear on a fulfillment queue with a guided pick-pack-ship workflow so that fulfillment is accurate and tracked.
- Constraint: Each item scanned during pick. Serial/lot numbers captured. Carrier and tracking entered at ship. Shopify order updated with tracking automatically.
Epic 4.J: Offline Operations
(New)
Story 4.J.1: Queue Inventory Changes Offline
- As a Store Staff Member, I want all inventory changes (sales, returns, adjustments, counts) to be queued locally during network outage so that I can continue working without interruption.
- Constraint: Local cache updated immediately. Maximum 500 queued changes. Queue warning at 90% capacity.
Story 4.J.2: Conflict Resolution on Reconnect
- As a Store Manager, I want the system to reconcile offline changes with server state on reconnect so that inventory accuracy is maintained across all locations.
- Constraint: Changes uploaded in chronological order. Negative inventory conflicts flagged for manager review. Last-write-wins for quantities by default.
Epic 4.K: Alerts & Notifications
(New)
Story 4.K.1: Configurable Inventory Alerts
- As an HQ Manager, I want configurable alerts for low stock, overstock, shrinkage, aging inventory, and overdue POs so that I am proactively informed of conditions requiring action.
- Constraint: Each alert type has configurable thresholds, severity, and delivery channels. Alert thresholds can be set per product per location.
Story 4.K.2: Email Templates
- As a Buyer, I want automated email templates for PO submission, transfer notifications, low stock digests, and count reminders so that stakeholders receive timely, formatted communications.
- Constraint: Four templates with dynamic field substitution. Templates support HTML formatting. Digest emails consolidate multiple alerts.
Story 4.K.3: Alert Acknowledgment
- As a Store Manager, I want to acknowledge alerts on the dashboard so that my team knows I am aware of and working on the issue.
- Constraint: Alert lifecycle: TRIGGERED > ACKNOWLEDGED > RESOLVED. Alerts auto-resolve when the triggering condition clears (except shrinkage, which requires manual resolution).
Epic 4.L: Dashboard & Reporting
(New)
Story 4.L.1: Dedicated Inventory Dashboard
- As a Store Manager, I want a dedicated inventory dashboard with KPI cards so that I can see the health of my store’s inventory at a glance.
- Constraint: Eight KPI cards (total value, low stock count, pending POs, open transfers, upcoming counts, shrinkage %, dead stock, avg days of supply). Filterable by location, category, brand.
Story 4.L.2: Analytics Report Suite
- As an HQ Manager, I want a comprehensive suite of 33 inventory reports so that I can analyze every aspect of inventory performance.
- Constraint: Reports exportable to CSV and PDF. Role-based access. Store managers see only their store’s data unless granted multi-store access.
Story 4.L.3: ABC Classification
- As a Buyer, I want monthly Pareto analysis that classifies products as A (top 20% revenue), B (next 30%), or C (bottom 50%) so that I can prioritize inventory management efforts.
- Constraint: New products exempt until 60 days of sales data. Classification drives cycle count frequency (A = 30 days, B = 60 days, C = 90 days).
Epic 4.M: PO Approval Workflow
(New)
Story 4.M.1: Threshold-Based PO Approval
- As an Owner, I want purchase orders above a configurable dollar threshold to require approval before submission so that large purchases are reviewed before committing funds.
- Constraint: Two tiers – manager approval at $500+, admin/owner approval at $5,000+. POs below $500 auto-approve.
Story 4.M.2: Auto-Approve Below Threshold
- As a Buyer, I want POs below the approval threshold to be submitted directly to the vendor without waiting for approval so that routine restocking is not delayed.
- Constraint: Auto-approved POs are still logged in the audit trail with source “AUTO_APPROVED”. Notification sent to manager for visibility.
Story 4.M.3: Approval Expiry
- As a Buyer, I want pending approvals to expire after a configurable period so that stale PO requests do not block the workflow indefinitely.
- Constraint: Default expiry is 7 days. Expired approvals notify the requester. Requester can resubmit.
Epic 4.N: Dead Stock Management
(New)
Story 4.N.1: Dead Stock Detection
- As an Inventory Manager, I want the system to automatically identify products with zero sales velocity over a configurable period so that I can take action on non-performing inventory.
- Constraint: Default threshold is 90 days of zero sales. Detection runs daily. Products with qty > 0 and velocity = 0 are flagged.
Story 4.N.2: Dead Stock Alerting
- As a Store Manager, I want to receive dashboard alerts for dead stock items so that I am aware of products that need markdowns, transfers, or vendor returns.
- Constraint: Alert includes product name, location, qty on hand, value (at WAC), and last sale date. Grouped by category.
Story 4.N.3: Dead Stock Reporting
- As a Buyer, I want a dead stock report showing all zero-velocity items with their value and age so that I can make informed decisions about clearance, vendor returns, or donation.
- Constraint: Report filterable by location, category, vendor, and value range. Exportable to CSV and PDF.
Epic 4.P: Overstock Vendor Returns
(New)
Story 4.P.1: Negotiate Vendor Return
- As a Buyer, I want to create overstock vendor return requests with proposed quantities and negotiate terms with the vendor so that excess inventory can be returned before it becomes dead stock.
- Constraint: Overstock returns follow the RMA workflow (Section 4.7). Must be enabled in config (
overstock_returns_enabled: true). Return window configurable per vendor.
Story 4.P.2: Restocking Fee Handling
- As a Buyer, I want to record restocking fees charged by the vendor so that the net credit is accurately tracked.
- Constraint: Restocking fee recorded as a percentage of unit cost. Default is 0%. Maximum configurable (default max: 25%). Net credit = (qty x cost) - restocking fee.
Story 4.P.3: Overstock Return Reporting
- As an HQ Manager, I want reports showing overstock return volume, restocking fees paid, and net credits recovered by vendor so that I can evaluate vendor return programs.
- Constraint: Includes vendor return rate, total credits recovered, total restocking fees, net recovery ratio. Filterable by vendor, date range, and category.
Inventory Acceptance Criteria: Gherkin Scenarios
Feature: Receiving with Discrepancy
Feature: Receiving with Discrepancy
As a Warehouse Clerk
I need to handle partial shipments and damaged goods during receiving
So that inventory records accurately reflect what was received
Background:
Given a Purchase Order "PO-2026-00142" exists in status "OPEN"
And the PO contains the following lines:
| SKU | Product | Ordered Qty | Unit Cost |
| SKU-1001 | Classic Fit Tee Navy M | 20 | $12.50 |
| SKU-1002 | Classic Fit Tee Navy L | 15 | $12.50 |
| SKU-1003 | Slim Chino Khaki 32 | 10 | $24.00 |
And the receiving location is "HQ Warehouse"
Scenario: Partial receive with all items in good condition
When I receive the following quantities:
| SKU | Received Qty |
| SKU-1001 | 20 |
| SKU-1002 | 10 |
And I do not receive SKU-1003
Then the stock level for "SKU-1001" at "HQ Warehouse" should increase by 20
And the stock level for "SKU-1002" at "HQ Warehouse" should increase by 10
And the stock level for "SKU-1003" should remain unchanged
And the PO status should update to "PARTIAL"
And the remaining qty for "SKU-1002" should be 5
And the remaining qty for "SKU-1003" should be 10
And WAC for "SKU-1001" should be recalculated using $12.50
And WAC for "SKU-1002" should be recalculated using $12.50
Scenario: Receive with damaged items quarantined
When I receive 20 units of "SKU-1001"
And I mark 3 units of "SKU-1001" as "DAMAGED" with reason "Stained in transit"
Then the stock level for "SKU-1001" at "HQ Warehouse" should show:
| Status | Qty |
| AVAILABLE | 17 |
| QUARANTINE | 3 |
And a movement log entry should be created with type "RECEIVE" and qty 20
And an RMA should be auto-created for 3 units of "SKU-1001"
And the RMA should reference "PO-2026-00142"
And the RMA reason should be "Stained in transit"
Scenario: Over-receive within threshold
Given the over-receive threshold is 10%
When I receive 22 units of "SKU-1001" (PO ordered 20)
Then the system should accept the over-receive (10% = 2 units, within threshold)
And the stock level for "SKU-1001" should increase by 22
And the PO line received qty should show 22
And an over-receive note should be logged
Scenario: Over-receive exceeds threshold requires approval
Given the over-receive threshold is 10%
When I attempt to receive 25 units of "SKU-1001" (PO ordered 20)
Then the system should block the receive with message "Over-receive of 25% exceeds 10% threshold"
And the system should prompt "Manager approval required for over-receive"
When manager "Mike" approves the over-receive
Then the stock level for "SKU-1001" should increase by 25
And the approval should be logged to the audit trail
Feature: POS Inventory Reservation
Feature: POS Inventory Reservation
As a POS system
I need to manage inventory reservations throughout the sale lifecycle
So that multiple terminals do not oversell available stock
Background:
Given product "Classic Fit Tee Navy M" (SKU-1001) has 5 units available at "Store A"
And Terminal 1 and Terminal 2 are active at "Store A"
Scenario: Reserve on add to cart
When cashier on Terminal 1 adds 1 unit of "SKU-1001" to the cart
Then a soft reservation should be created for Terminal 1
And available qty for "SKU-1001" at "Store A" should show 4
And Terminal 2 should see available qty as 4
Scenario: Release on item removal
Given Terminal 1 has 1 unit of "SKU-1001" in the cart (reserved)
And available qty shows 4
When cashier on Terminal 1 removes "SKU-1001" from the cart
Then the soft reservation should be released
And available qty for "SKU-1001" at "Store A" should show 5
Scenario: Commit on successful payment
Given Terminal 1 has 1 unit of "SKU-1001" in the cart (reserved)
When payment completes successfully on Terminal 1
Then the soft reservation should be deleted
And a SALE movement should be logged with qty -1
And available qty for "SKU-1001" at "Store A" should show 4 (permanent)
And WAC should be captured on the movement record
Scenario: Hold on payment failure then auto-release
Given Terminal 1 has 1 unit of "SKU-1001" in the cart (reserved)
And available qty shows 4
When payment fails on Terminal 1 (card declined)
Then the reservation should hold for 30 seconds
And available qty should remain 4 during the hold
When 30 seconds elapse without payment retry
Then the reservation should auto-release
And available qty should return to 5
Scenario: Hold on payment failure with retry
Given Terminal 1 has 1 unit of "SKU-1001" in the cart (reserved)
When payment fails on Terminal 1 (card declined)
And cashier retries payment within 30 seconds
And the retry succeeds
Then the existing reservation should be used (no new reservation)
And a SALE movement should be logged
And available qty should show 4 (permanent)
Scenario: Void releases all reservations
Given Terminal 1 has 2 units of "SKU-1001" in the cart (reserved)
And available qty shows 3
When cashier voids the entire transaction
Then all reservations for the session should be released
And available qty for "SKU-1001" should return to 5
And no movement records should be created
Scenario: Last unit contention between terminals
Given available qty for "SKU-1001" is 1
When cashier on Terminal 1 adds 1 unit of "SKU-1001" to cart
Then available qty should show 0
When cashier on Terminal 2 attempts to add "SKU-1001" to cart
Then Terminal 2 should see "SKU-1001 is out of stock at this location"
And the item should not be added to Terminal 2's cart
Feature: Inventory Count with Scanner
Feature: Inventory Count with Scanner
As a Store Manager
I need scanner-primary counting with variance detection
So that counts are accurate and discrepancies are reviewed
Background:
Given a cycle count session is created for category "Tops" at "Store A"
And the following expected quantities exist:
| SKU | Product | Expected Qty |
| SKU-1001 | Classic Fit Tee Navy M | 25 |
| SKU-1002 | Classic Fit Tee Navy L | 18 |
| SKU-1003 | V-Neck Tee Black M | 12 |
And blind count mode is enabled (expected qty hidden from counter)
And scanner mode is "scan_primary"
And variance approval threshold is 10 units
Scenario: Scanner-primary count with no variance
When staff scans 25 barcodes for "SKU-1001"
And staff scans 18 barcodes for "SKU-1002"
And staff scans 12 barcodes for "SKU-1003"
And staff submits the count
Then all items should show zero variance
And the count status should change to "COMPLETED"
And no adjustments should be created
And the count should be logged to history
Scenario: Count with minor variance auto-adjusts
When staff scans 23 barcodes for "SKU-1001" (expected 25)
And staff scans 18 barcodes for "SKU-1002"
And staff scans 12 barcodes for "SKU-1003"
And staff submits the count
Then variance for "SKU-1001" should show -2 units
And the variance is below the 10-unit threshold
Then "SKU-1001" inventory should be auto-adjusted to 23
And an adjustment record should be created with reason "COUNT_CORRECTION"
And the movement log should record qty change of -2 for "SKU-1001"
Scenario: Count with major variance requires approval
When staff scans 12 barcodes for "SKU-1001" (expected 25)
And staff submits the count
Then variance for "SKU-1001" should show -13 units
And the variance exceeds the 10-unit threshold
Then the adjustment should be set to "PENDING_APPROVAL" status
And manager should receive an approval notification
And "SKU-1001" inventory should remain at 25 until approved
Scenario: Manager approves high-variance adjustment
Given a pending adjustment exists for "SKU-1001" with counted qty 12 (variance -13)
When manager "Mike" reviews and approves the adjustment
Then "SKU-1001" inventory should be adjusted to 12
And the adjustment reason should be "COUNT_CORRECTION"
And the approval should be logged with approver "Mike"
And a SHRINKAGE alert should be triggered (variance 52% exceeds 5% threshold)
Scenario: Recount required for extreme variance
Given variance approval threshold for recount is 20%
When staff counts 5 units for "SKU-1001" (expected 25, variance 80%)
And staff submits the count
Then the system should flag "SKU-1001" for mandatory recount
And the count status for "SKU-1001" should be "RECOUNT_REQUIRED"
And a different staff member should be assigned the recount
Feature: PO Approval Workflow
Feature: PO Approval Workflow
As an Owner
I need purchase orders above a dollar threshold to require approval
So that large purchases are reviewed before funds are committed
Background:
Given the following PO approval thresholds are configured:
| Threshold | Required Approver |
| < $500 | Auto-approve |
| >= $500 | Manager |
| >= $5000 | Admin/Owner |
And approval requests expire after 7 days
Scenario: Auto-approve PO below threshold
When buyer "Sarah" creates PO "PO-2026-00201" with total value $350.00
And submits the PO
Then the PO should be auto-approved
And the PO status should change to "OPEN"
And the audit log should record source "AUTO_APPROVED"
And an email should be sent to the vendor
And manager "Mike" should receive a visibility notification
Scenario: Manager approval required
When buyer "Sarah" creates PO "PO-2026-00202" with total value $1,200.00
And submits the PO for approval
Then an approval request should be created with status "PENDING"
And the PO status should remain "DRAFT"
And manager "Mike" should receive an approval notification
Scenario: Manager approves PO
Given a pending approval exists for PO "PO-2026-00202" ($1,200.00)
When manager "Mike" approves the PO
Then the PO status should change to "OPEN"
And the vendor should receive the PO email (TMPL_INV_PO_VENDOR)
And the audit log should record approver "Mike"
Scenario: Manager rejects PO
Given a pending approval exists for PO "PO-2026-00202" ($1,200.00)
When manager "Mike" rejects the PO with reason "Wait for vendor sale next month"
Then the PO status should remain "DRAFT"
And buyer "Sarah" should receive a notification with the rejection reason
And the rejection should be logged to the audit trail
Scenario: Large PO escalates to admin
When buyer "Sarah" creates PO "PO-2026-00203" with total value $7,500.00
And submits the PO for approval
Then an approval request should require "ADMIN" level approval
And manager approval should NOT be sufficient
And admin/owner should receive the approval notification
Scenario: Approval request expires
Given a pending approval was created 8 days ago for PO "PO-2026-00204"
When the expiration job runs
Then the approval status should change to "EXPIRED"
And the PO status should remain "DRAFT"
And buyer "Sarah" should be notified of the expiration
And "Sarah" should be able to resubmit the PO for approval
Scenario: Requester cannot approve own PO
Given buyer "Sarah" also has Manager role
When "Sarah" creates and submits PO "PO-2026-00205" ($800.00)
Then "Sarah" should not be able to approve her own PO
And the system should show "Cannot approve your own purchase order"
And a different manager must approve
Feature: Transfer Auto-Suggestion
Feature: Transfer Auto-Suggestion
As an HQ Manager
I need the system to detect inventory imbalances and suggest transfers
So that stock is distributed optimally across stores
Background:
Given the auto-suggest imbalance threshold is 14 days of supply
And minimum transfer quantity is 2 units
And minimum source qty after transfer is 2 units
And the following inventory state exists:
| Product | Store A (qty/velocity) | Store B (qty/velocity) | Store C (qty/velocity) |
| Classic Tee Navy | 30 / 1.0 per day | 2 / 1.5 per day | 15 / 0.5 per day |
And days of supply:
| Product | Store A | Store B | Store C |
| Classic Tee Navy | 30 days | 1.3 days| 30 days |
Scenario: Imbalance detected and suggestion generated
When the daily auto-suggest job runs
Then a transfer suggestion should be generated
And the suggestion should recommend moving stock from "Store A" to "Store B"
And the suggested qty should equalize days of supply across stores
And the suggestion should not reduce "Store A" below minimum qty (2 units)
Scenario: Manager reviews and approves suggestion
Given a transfer suggestion exists: 10 units of "Classic Tee Navy" from "Store A" to "Store B"
When HQ manager "Alex" reviews the suggestion
And approves the transfer
Then a transfer record should be created in "APPROVED" status
And "Store A" staff should be notified to pick and ship 10 units
And the suggestion status should change to "ACCEPTED"
Scenario: Manager modifies suggestion before approval
Given a transfer suggestion exists: 10 units from "Store A" to "Store B"
When HQ manager "Alex" changes the quantity to 8 units
And approves the modified transfer
Then a transfer record should be created for 8 units
And the suggestion should be marked "ACCEPTED_MODIFIED"
Scenario: Manager rejects suggestion
Given a transfer suggestion exists: 10 units from "Store A" to "Store B"
When HQ manager "Alex" rejects the suggestion with reason "Seasonal event at Store A"
Then no transfer should be created
And the suggestion should be marked "REJECTED" with the reason
Scenario: No suggestion when imbalance below threshold
Given all stores have days of supply within 14 days of each other
When the daily auto-suggest job runs
Then no transfer suggestions should be generated
Feature: Offline Inventory Sync
Feature: Offline Inventory Sync
As a POS system
I need to queue inventory changes offline and reconcile on reconnect
So that stores can continue operating during network outages
Background:
Given "Store A" has the following inventory:
| SKU | Available Qty |
| SKU-1001 | 10 |
| SKU-2005 | 5 |
And the conflict resolution strategy is "last_write_wins"
And maximum queue size is 500
Scenario: Queue sales offline and sync on reconnect
Given "Store A" loses network connectivity
And POS enters OFFLINE mode
When cashier sells 2 units of "SKU-1001" at 10:15 AM
And cashier sells 1 unit of "SKU-2005" at 10:30 AM
Then local cache should show:
| SKU | Available Qty |
| SKU-1001 | 8 |
| SKU-2005 | 4 |
And the offline queue should contain 2 entries
When network connectivity restores
Then POS should enter SYNCING mode
And all queued changes should upload in chronological order
And server should apply: SKU-1001 qty -2, SKU-2005 qty -1
And POS should enter ONLINE mode
And queue should be empty
Scenario: Conflict detected on reconnect (negative inventory)
Given "Store A" loses network connectivity
And while offline, Store A cashier sells 4 units of "SKU-2005" (local cache: 1)
And while offline, Store B sells 3 units of "SKU-2005" on the server (server qty: 2)
When "Store A" reconnects and syncs the offline sale of 4 units
Then server calculates: server qty (2) - offline change (4) = -2
And a conflict should be flagged:
| Field | Value |
| product | SKU-2005 |
| offline_change | -4 |
| server_qty_at_sync | 2 |
| resulting_qty | -2 |
| conflict_type | NEGATIVE_INVENTORY |
And POS should enter CONFLICT_REVIEW mode
And manager should be notified of the conflict
Scenario: Manager resolves negative inventory conflict
Given a conflict exists for "SKU-2005" with resulting qty -2
When manager "Mike" reviews the conflict
And accepts the negative balance with note "Schedule recount"
Then the adjustment should be applied (qty = -2)
And a recount should be scheduled for "SKU-2005" at "Store A"
And the conflict status should change to "ACCEPTED"
And POS should return to ONLINE mode
Scenario: Blocked operations show clear messages
Given "Store A" is in OFFLINE mode
When staff attempts to check inventory at other stores
Then the system should display "Multi-store lookup unavailable offline. Check local stock only."
When staff attempts to create a transfer request
Then the system should display "Transfers unavailable offline. Queue request when online."
When staff attempts to receive a PO
Then the system should display "Receiving unavailable offline. Wait for connectivity."
Scenario: Queue capacity warning
Given "Store A" is in OFFLINE mode
And the offline queue has 450 of 500 entries (90%)
Then POS should display warning "Offline queue nearly full. Reconnect soon."
When the queue reaches 500 entries
Then POS should block further inventory-modifying operations
And display "Offline queue full. Cannot process more transactions until reconnected."
Feature: Dead Stock Detection
Feature: Dead Stock Detection
As an Inventory Manager
I need to identify products with zero sales velocity
So that dead stock can be addressed through markdowns, transfers, or vendor returns
Background:
Given the dead stock threshold is 90 days
And the following inventory exists at "Store A":
| SKU | Product | Qty | Last Sale Date |
| SKU-3001 | Printed Scarf | 15 | 2025-10-01 |
| SKU-3002 | Wool Beanie | 8 | 2026-01-15 |
| SKU-3003 | Linen Blazer | 3 | 2025-08-20 |
And today's date is 2026-02-04
Scenario: Product flagged as dead stock after 90 days
When the daily dead stock detection job runs
Then "SKU-3001" should be flagged as dead stock (126 days since last sale)
And "SKU-3003" should be flagged as dead stock (168 days since last sale)
And "SKU-3002" should NOT be flagged (20 days since last sale)
Scenario: Dead stock alert created
When "SKU-3001" is flagged as dead stock
Then an AGING_INVENTORY alert should be created with severity "WARNING"
And the alert data snapshot should include:
| Field | Value |
| available_qty | 15 |
| last_sale_date | 2025-10-01 |
| days_since_sale | 126 |
| value_at_wac | calculated |
Scenario: Dead stock report generated
When manager requests the Dead Stock Report for "Store A"
Then the report should include "SKU-3001" and "SKU-3003"
And the report should NOT include "SKU-3002"
And each row should show: SKU, product name, qty, WAC value, last sale date, days since sale
And the report should be exportable to CSV and PDF
Scenario: Dead stock auto-resolves when sale occurs
Given "SKU-3001" has an active AGING_INVENTORY alert
When a customer purchases 1 unit of "SKU-3001" at "Store A"
Then the AGING_INVENTORY alert for "SKU-3001" should auto-resolve
And the resolved_at timestamp should be set
And auto_resolved should be true
Feature: Online Order Fulfillment
Feature: Online Order Fulfillment
As an Online Operations Manager
I need online orders to be assigned to the nearest store and fulfilled
So that customers receive orders quickly and shipping costs are minimized
Background:
Given the following stores with locations:
| Store | City | Lat | Lng | SKU-1001 Qty |
| Store A | Richmond, VA | 37.54 | -77.44 | 10 |
| Store B | Norfolk, VA | 36.85 | -76.29 | 0 |
| Store C | Virginia Beach| 36.85 | -75.98 | 5 |
| HQ | Glen Allen, VA| 37.66 | -77.51 | 25 |
And store assignment strategy is "nearest"
And split fulfillment is disabled
Scenario: Nearest store with stock is selected
Given a customer in "Chesapeake, VA" (lat 36.77, lng -76.29) places an online order for 2 units of "SKU-1001"
When the store assignment algorithm runs
Then "Store C" should be selected (nearest with stock, ~22 miles)
And 2 units of "SKU-1001" should be hard-reserved at "Store C"
And the order should appear on "Store C" fulfillment queue
Scenario: Nearest store has no stock, next nearest selected
Given "Store C" has 0 units of "SKU-1001" (sold out)
And a customer in "Chesapeake, VA" places an online order for 2 units of "SKU-1001"
When the store assignment algorithm runs
Then "Store A" should be selected (next nearest with stock)
And 2 units should be hard-reserved at "Store A"
Scenario: No store has stock, order flagged for manual assignment
Given all stores have 0 units of "SKU-1001"
And HQ has 25 units of "SKU-1001"
And exclude_hq is false
When the store assignment algorithm runs
Then "HQ" should be selected for fulfillment
And 2 units should be hard-reserved at "HQ"
Scenario: Pick-pack-ship workflow completes
Given order "ORD-SHOP-9001" is assigned to "Store A"
And the order contains 2 units of "SKU-1001"
When staff starts picking
And scans 2 barcodes for "SKU-1001"
Then picking should be complete
When staff packs the order and enters weight
And enters carrier "UPS" with tracking "1Z999AA10123456784"
Then the order status should be "SHIPPED"
And a SALE movement should be logged for 2 units at "Store A"
And Shopify should be updated with fulfillment and tracking number
And available qty for "SKU-1001" at "Store A" should decrease by 2
Scenario: Inventory sync from Shopify online sale
Given "Store A" has 10 units of "SKU-1001" in POS
And Shopify shows 10 units for "Store A"
When a customer purchases 1 unit online (assigned to Store A)
Then Shopify webhook fires orders/create
And POS should decrement "SKU-1001" at "Store A" by 1
And POS should show 9 units available
And the movement type should be "ONLINE_SALE"
Feature: Overstock Vendor Return
Feature: Overstock Vendor Return
As a Buyer
I need to return overstock items to vendors with restocking fee tracking
So that excess inventory can be managed and credits recovered
Background:
Given overstock returns are enabled
And the default restocking fee is 0%
And vendor "StyleCo" has a negotiated restocking fee of 15%
And the following overstock inventory exists:
| SKU | Product | Qty | Days of Supply | Vendor |
| SKU-4001 | Floral Dress S | 50 | 250 days | StyleCo |
| SKU-4002 | Floral Dress M | 35 | 175 days | StyleCo |
Scenario: Create overstock vendor return
When buyer "Sarah" creates an overstock RMA for vendor "StyleCo"
And adds the following return lines:
| SKU | Return Qty | Unit Cost |
| SKU-4001 | 30 | $18.00 |
| SKU-4002 | 20 | $18.00 |
Then an RMA should be created with type "OVERSTOCK"
And the RMA status should be "DRAFT"
And the gross return value should be $900.00
Scenario: Restocking fee applied on vendor approval
Given RMA "RMA-2026-00050" is submitted to vendor "StyleCo"
When vendor approves the return with 15% restocking fee
Then the restocking fee should be $135.00 (15% of $900.00)
And the net credit expected should be $765.00
And the RMA status should change to "VENDOR_APPROVED"
And the restocking fee should be recorded on each line:
| SKU | Gross Credit | Restocking Fee | Net Credit |
| SKU-4001 | $540.00 | $81.00 | $459.00 |
| SKU-4002 | $360.00 | $54.00 | $306.00 |
Scenario: Inventory decremented on shipment back to vendor
Given RMA "RMA-2026-00050" is vendor-approved
When warehouse ships the return items back to vendor
And enters carrier "FedEx" with tracking "794644790132"
Then the RMA status should change to "SHIPPED_BACK"
And a RETURN_TO_VENDOR movement should be logged:
| SKU | Qty Change | Location |
| SKU-4001 | -30 | HQ Warehouse |
| SKU-4002 | -20 | HQ Warehouse |
And inventory at "HQ Warehouse" should decrease accordingly
Scenario: Credit received and reconciled
Given RMA "RMA-2026-00050" was shipped back to vendor
When vendor issues credit of $765.00
And buyer records the credit received
Then the RMA status should change to "CREDIT_RECEIVED"
And credit reconciliation should show:
| Expected Credit | Received Credit | Variance |
| $765.00 | $765.00 | $0.00 |
And the RMA status should change to "CLOSED"
Scenario: Credit variance detected
Given RMA "RMA-2026-00050" was shipped back with expected credit $765.00
When vendor issues credit of $700.00 (less than expected)
And buyer records the credit received
Then credit reconciliation should show variance of -$65.00
And a reconciliation alert should be created
And buyer should investigate the discrepancy before closing the RMA
Feature: Inventory Adjustment Approval
Feature: Inventory Adjustment Approval
As a Store Manager
I need inventory adjustments above a threshold to require approval
So that significant inventory changes are reviewed for accuracy
Background:
Given adjustment approval mode is "threshold"
And the approval threshold is 10 units or $100.00 value
And product "SKU-1001" has 25 units available at "Store A"
And WAC for "SKU-1001" is $12.50
Scenario: Small adjustment auto-applies
When staff "Jane" submits an adjustment for "SKU-1001" at "Store A"
And the adjustment is -3 units (value: $37.50) with reason "DAMAGED"
Then the adjustment should be applied immediately
And "SKU-1001" qty at "Store A" should change to 22
And a movement record should be created:
| Type | Qty | Reason | Staff |
| ADJUSTMENT | -3 | DAMAGED | Jane |
And no approval request should be created
Scenario: Large adjustment requires approval
When staff "Jane" submits an adjustment for "SKU-1001" at "Store A"
And the adjustment is -15 units (value: $187.50) with reason "THEFT"
Then the adjustment should be set to "PENDING_APPROVAL" status
And "SKU-1001" qty should remain at 25 until approved
And manager "Mike" should receive an approval notification
And the notification should include: product, qty change, value, reason, requester
Scenario: Manager approves the adjustment
Given a pending adjustment exists: "SKU-1001" at "Store A", -15 units, reason "THEFT"
When manager "Mike" reviews and approves the adjustment
Then "SKU-1001" qty at "Store A" should change to 10
And a movement record should be created with approver "Mike"
And the adjustment status should be "APPROVED"
And staff "Jane" should be notified of the approval
Scenario: Manager rejects the adjustment
Given a pending adjustment exists: "SKU-1001" at "Store A", -15 units, reason "THEFT"
When manager "Mike" rejects the adjustment with reason "Recount needed first"
Then "SKU-1001" qty should remain at 25
And the adjustment status should be "REJECTED"
And staff "Jane" should be notified with the rejection reason
Scenario: Reason code is required
When staff "Jane" submits an adjustment for "SKU-1001" without selecting a reason code
Then the system should display "Reason code is required for inventory adjustments"
And the adjustment should be blocked
Scenario: Custom reason code with required note
When staff "Jane" submits an adjustment with reason code "OTHER"
And does not provide a note
Then the system should display "A note is required for reason code 'Other'"
And the adjustment should be blocked
When "Jane" provides note "Found box behind shelf during cleaning"
And resubmits
Then the adjustment should proceed (subject to threshold rules)
5. Setup & Configuration Module
5.1 Overview & Scope
Module 5 centralizes all tenant-level system configuration for the POS platform. Every operational behavior in Modules 1 through 4 – how a sale is processed, how a customer is identified, how a product is priced, how inventory is tracked – is governed by configuration defined here. Module 5 is the control plane: it does not process transactions, manage catalog records, or move inventory. It defines the rules, structures, and parameters that those modules consume at runtime.
5.1.1 Executive Summary
A multi-tenant POS system serving five retail stores and one HQ warehouse requires a single, authoritative source for all system-wide configuration. Without centralized setup, configuration drifts across locations, roles are inconsistently enforced, and operational rules become embedded in application logic rather than tenant-controlled settings. Module 5 eliminates this by providing a structured configuration layer that every other module references.
The scope of Module 5 encompasses: system identity and branding, currency and localization, physical location definitions, user identity and role-based access, shift and scheduling configuration, register and terminal management, hardware peripherals, tax rules, receipt templates, payment processing integration, financial accounting codes, operational business rules, inter-store transfer policies, notification and alert routing, RFID hardware integration, system integrations with external platforms, data import/export tooling, audit logging configuration, and tenant onboarding workflows.
Design principle: Module 5 defines how things are configured, not how things operate day-to-day. For example, Module 5 defines that a location exists, its timezone, and its tax rate. Module 1 (Sales) uses that tax rate when calculating a transaction total. Module 5 defines that a user has the MANAGER role with permission to void transactions. Module 1 enforces that permission at the point of sale.
5.1.2 Module Dependencies
Module 5 is the foundational configuration layer consumed by all operational modules. It has no upstream module dependencies – it is configured directly by tenant administrators.
flowchart TD
M5["Module 5\nSetup & Configuration"]
M1["Module 1\nSales & POS"]
M2["Module 2\nCustomers"]
M3["Module 3\nCatalog"]
M4["Module 4\nInventory"]
M5 -->|Users, roles, registers,\ntax rules, payment config,\nreceipt templates, clock-in/out config| M1
M5 -->|Users, roles,\nlocation assignments,\ncustomer data policies| M2
M5 -->|Users, roles,\nbarcode config,\nvendor settings,\nlabel printers| M3
M5 -->|Users, roles,\nlocations,\ntransfer rules,\nRFID config| M4
style M5 fill:#7b2d8e,stroke:#5a1d6e,color:#fff
style M1 fill:#264653,stroke:#1d3557,color:#fff
style M2 fill:#264653,stroke:#1d3557,color:#fff
style M3 fill:#264653,stroke:#1d3557,color:#fff
style M4 fill:#264653,stroke:#1d3557,color:#fff
Downstream consumers (Module 5 provides):
| Consumer Module | Configuration Provided | Purpose |
|---|---|---|
| Module 1 (Sales) | Registers, profiles, tax rates, payment processors, receipt templates, user roles, clock-in/out configuration, cash drawer settings, discount limits | Controls POS terminal behavior, payment routing, receipt output, and staff permissions during sales. |
| Module 2 (Customers) | User roles, data retention policies, communication preferences defaults, location assignments | Governs who can view/edit customer data, default privacy settings, and location-scoped customer association. |
| Module 3 (Catalog) | Barcode format, label printer config, vendor registry, user roles, approval thresholds, Shopify integration settings | Controls barcode generation, label printing, vendor management permissions, and external catalog sync. |
| Module 4 (Inventory) | Locations, transfer rules, RFID reader config, user roles, reorder thresholds, count policies | Defines physical topology, transfer approval rules, and counting schedules. |
5.1.3 Functional Scope
The following table enumerates all functional areas covered by Module 5 and their section references.
| # | Section | Domain | Description |
|---|---|---|---|
| 5.1 | Overview & Scope | Foundation | Module purpose, dependencies, and section index |
| 5.2 | System Settings & Branding | Identity | Tenant identity, operational defaults, and visual branding |
| 5.3 | Multi-Currency Configuration | Localization | Currency definitions, exchange rates, and display formatting |
| 5.4 | Locations | Topology | Physical locations and location type definitions |
| 5.5 | Users & Roles | Access Control | User profiles, role definitions, and feature toggle matrix |
| 5.6 | Time Tracking (Clock-In / Clock-Out) | Time Tracking | Simple clock-in/clock-out time recording for payroll |
| 5.7 | Registers & Terminals | Hardware | Register registry, device pairing, profiles, and peripheral assignments |
| 5.8 | Printer Configuration | Hardware | Printer registry, driver settings, and printer-location assignments |
| 5.9 | Tax Configuration | Financial | Tax rates, tax classes, and location-level tax rules |
| 5.10 | Notification & Alert Rules | Communication | Alert routing, escalation paths, and notification channel preferences |
| 5.11 | Payment Processing | Financial | Payment processor integration, terminal pairing, and gateway configuration |
| 5.12 | Accounting & GL Mapping | Financial | Chart of accounts, GL code assignments, and financial period definitions |
| 5.13 | Operational Business Rules | Rules Engine | Configurable thresholds, approval limits, and policy toggles |
| 5.14 | Receipt Templates | Output | Receipt layout configuration, template variables, and format options |
| 5.15 | Transfer & Logistics Rules | Operations | Transfer approval thresholds, routing rules, and carrier configuration |
| 5.16 | RFID Configuration | Hardware | Reader registration, antenna settings, and scan session parameters |
| 5.17 | System Integrations | External | Shopify, QuickBooks, and third-party API connection management |
| 5.18 | Data Import / Export | Data | Bulk data import templates, export scheduling, and format configuration |
| 5.19 | Audit & Compliance | Governance | Audit log retention, compliance settings, and data purge policies |
| 5.20 | Tenant Onboarding | Lifecycle | Initial setup wizard, seed data provisioning, and go-live checklist |
| 5.21 | User Stories | Acceptance | Gherkin-format acceptance criteria for all Module 5 functionality |
5.2 System Settings & Branding
Scope: Tenant-level identity, operational defaults, and visual branding configuration. These settings establish the foundational parameters that all other modules reference – the tenant’s name, timezone, date formatting, session policies, and customer-facing visual identity. Settings are organized into three categories: Core (identity and localization), Operational (runtime behavior), and Branding (visual presentation).
5.2.1 Core Settings
Core settings define the tenant’s identity and localization defaults. These values are established during onboarding (Section 5.20) and rarely change after initial configuration.
| Setting | Key | Type | Required | Default | Description |
|---|---|---|---|---|---|
| Tenant Name | tenant_name | String(100) | Yes | – | Trading name displayed in UI headers and reports |
| Legal Entity Name | legal_entity_name | String(200) | Yes | – | Registered business name for invoices and legal documents |
| Company Logo | company_logo_url | String(500) | No | System default | URL to uploaded logo image (PNG/SVG, max 2MB, min 200x200px) |
| Default Timezone | default_timezone | String(50) | Yes | America/New_York | IANA timezone identifier; applies to all locations unless overridden at location level |
| Default Currency | default_currency | String(3) | Yes | USD | ISO 4217 currency code; set at onboarding, immutable after first transaction |
| Date Format | date_format | Enum | Yes | MM/DD/YYYY | Display format: MM/DD/YYYY or DD/MM/YYYY; tenant-wide preference |
| Time Format | time_format | Enum | Yes | 12h | Display format: 12h (3:00 PM) or 24h (15:00); tenant-wide preference |
| Fiscal Year Start | fiscal_year_start_month | Integer(1-12) | Yes | 1 (January) | Month number when the fiscal year begins; affects financial reporting periods |
5.2.2 Operational Settings
Operational settings control runtime behavior across the POS system. These are actively tuned by administrators as business needs evolve.
| Setting | Key | Type | Required | Default | Description |
|---|---|---|---|---|---|
| Auto-Logout Timeout | auto_logout_minutes | Integer | Yes | 30 | Minutes of inactivity before automatic POS session logout (range: 5-120) |
| Max Session Duration | max_session_hours | Integer | Yes | 8 | Maximum hours a POS session can remain active before forced re-authentication (range: 1-24) |
| Barcode Format | barcode_format | Enum | Yes | CODE128 | Preferred barcode symbology for system-generated barcodes: CODE128, EAN13, UPCA |
| Default Print Mode | default_print_mode | Enum | Yes | THERMAL | Default receipt output: THERMAL (80mm roll), A4 (full page), EMAIL_ONLY |
| Failed Login Lockout | failed_login_max | Integer | Yes | 5 | Number of consecutive failed PIN/password attempts before temporary lockout |
| Lockout Duration | lockout_duration_minutes | Integer | Yes | 15 | Minutes a user account is locked after exceeding failed login threshold |
Cross-Reference: See Module 5, Section 5.14 for receipt template configuration including default printer assignment per location.
5.2.3 Business Hours & Holiday Calendar
Business hours are configured per location, supporting different schedules per day of week. Holidays override normal business hours for specific dates.
Business Hours Data Model
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
location_id | UUID | Yes | – | FK to locations table |
day_of_week | Integer(0-6) | Yes | – | 0 = Sunday, 1 = Monday, …, 6 = Saturday |
open_time | Time | No | 09:00 | Store opening time (null = closed this day) |
close_time | Time | No | 21:00 | Store closing time (null = closed this day) |
is_closed | Boolean | Yes | false | Overrides open/close times; true = location closed this day |
Holiday Calendar Data Model
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
holiday_date | Date | Yes | – | Calendar date of the holiday |
name | String(100) | Yes | – | Holiday name (e.g., “Thanksgiving”, “Independence Day”) |
applies_to | Enum | Yes | ALL | ALL (all locations), STORES_ONLY, SPECIFIC (use junction table) |
is_closed | Boolean | Yes | true | Whether locations are closed on this date |
modified_open_time | Time | No | null | Override open time if not fully closed (e.g., holiday shortened hours) |
modified_close_time | Time | No | null | Override close time if not fully closed |
is_recurring | Boolean | Yes | false | If true, repeats annually on the same month/day |
Business Rules:
- Holiday entries override normal business hours for matching dates.
- When
is_closed = true, POS terminals at affected locations display a “Location Closed” banner and block new transactions. - When modified hours are set, the system uses those hours instead of normal business hours for that date.
- Recurring holidays auto-generate entries for the current fiscal year during onboarding and can be manually adjusted per year.
5.2.4 Branding Settings
Branding settings control the visual presentation of the POS system, customer-facing displays, printed receipts, and exported reports.
| Setting | Key | Type | Required | Default | Description |
|---|---|---|---|---|---|
| Primary Color | brand_primary_color | String(7) | Yes | #1A1A2E | Hex color code for primary UI elements (headers, buttons, navigation) |
| Accent Color | brand_accent_color | String(7) | Yes | #E94560 | Hex color code for accent elements (highlights, active states, badges) |
| Login Background | login_bg_image_url | String(500) | No | System default | URL to custom login page background image (JPG/PNG, max 5MB, 1920x1080 recommended) |
| Login Tagline | login_tagline | String(200) | No | null | Custom text displayed below the company logo on the login screen |
| Receipt Logo | receipt_logo_url | String(500) | No | company_logo_url | Logo printed/displayed on receipts; falls back to company logo if not set |
| Report Header Logo | report_header_logo_url | String(500) | No | company_logo_url | Logo displayed in the header of printed and exported reports |
| Report Header Address | report_header_address | Text | No | null | Company address block printed on report headers |
| Customer Display Welcome | customer_display_welcome | String(200) | No | "Welcome!" | Welcome message shown on customer-facing displays |
| Customer Display Promo Images | customer_display_promo_urls | JSONB | No | [] | Array of image URLs for rotating promotional display on customer-facing screens |
Cross-Reference: See Module 5, Section 5.14 for receipt-specific branding (receipt logo placement, footer text, color printing support).
5.2.5 System Settings Data Model
All settings are stored in a key-value table with JSONB values, enabling flexible schema evolution without database migrations.
system_settings Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
setting_key | String(100) | Yes | – | Unique setting identifier within tenant (e.g., tenant_name, auto_logout_minutes) |
setting_value | JSONB | Yes | – | Setting value; JSONB supports strings, numbers, booleans, arrays, and objects |
category | Enum | Yes | – | CORE, OPERATIONAL, BRANDING |
updated_by | UUID | Yes | – | FK to users table; last user who modified this setting |
updated_at | DateTime | Yes | Auto | Timestamp of last modification |
Unique constraint: (tenant_id, setting_key)
location_settings Table (Per-Location Overrides)
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
location_id | UUID | Yes | – | FK to locations table |
setting_key | String(100) | Yes | – | Setting identifier being overridden at location level |
setting_value | JSONB | Yes | – | Location-specific override value |
updated_by | UUID | Yes | – | FK to users table |
updated_at | DateTime | Yes | Auto | Timestamp of last modification |
Unique constraint: (tenant_id, location_id, setting_key)
Resolution order: Location-level setting overrides tenant-level setting. If no location override exists, the tenant-level value is used.
Overridable settings: Not all settings support location-level override. The following settings are overridable per location: default_timezone, auto_logout_minutes, max_session_hours, default_print_mode, barcode_format.
5.3 Multi-Currency Configuration
Scope: Defining the base currency and optional additional currencies for vendor purchase order support. All POS sales transactions and financial reports operate exclusively in the tenant’s base currency. Multi-currency support exists solely to enable purchase orders denominated in a vendor’s native currency, with manual exchange rate management and date-stamped rate history.
5.3.1 Currency Model
Each tenant has exactly one base currency, established at onboarding and immutable after the first transaction is recorded. Additional currencies are activated to support vendor procurement workflows where the vendor invoices in a foreign currency.
Currency Data Model
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
code | String(3) | Yes | – | ISO 4217 currency code (e.g., USD, EUR, GBP, CAD) |
name | String(50) | Yes | – | Display name (e.g., “US Dollar”, “Euro”, “British Pound”) |
symbol | String(5) | Yes | – | Currency symbol (e.g., $, €, £) |
decimal_places | Integer | Yes | 2 | Number of decimal places for amounts (0-4) |
symbol_position | Enum | Yes | BEFORE | Symbol placement: BEFORE ($100.00) or AFTER (100.00€) |
thousands_separator | String(1) | Yes | , | Thousands grouping character: , or . or (space) |
decimal_separator | String(1) | Yes | . | Decimal point character: . or , |
is_base | Boolean | Yes | false | Whether this is the tenant’s base currency (exactly one per tenant) |
is_active | Boolean | Yes | true | Whether this currency is available for selection on new POs |
created_at | DateTime | Yes | Auto | Record creation timestamp |
updated_at | DateTime | Yes | Auto | Last modification timestamp |
Unique constraint: (tenant_id, code)
Business Rules:
- Exactly one currency per tenant must have
is_base = true. This constraint is enforced at the application level. - The base currency cannot be deactivated (
is_activecannot be set tofalsefor the base currency). - The base currency code cannot be changed after the first transaction is recorded in the system.
- Deactivating a non-base currency prevents it from being selected on new POs but does not affect existing POs already denominated in that currency.
5.3.2 Exchange Rates
Exchange rates are manually entered by an administrator. The system does not integrate with external rate feeds or auto-update rates. Each rate entry is date-stamped; the system uses the most recent rate on or before the PO date when converting vendor currency amounts to the base currency.
Exchange Rate Data Model
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
from_currency_id | UUID | Yes | – | FK to currencies table; the source currency |
to_currency_id | UUID | Yes | – | FK to currencies table; the target currency (typically the base currency) |
rate | Decimal(12,6) | Yes | – | Exchange rate: 1 unit of from_currency = rate units of to_currency |
effective_date | Date | Yes | – | Date from which this rate is effective |
created_by | UUID | Yes | – | FK to users table; administrator who entered the rate |
created_at | DateTime | Yes | Auto | Record creation timestamp |
Unique constraint: (tenant_id, from_currency_id, to_currency_id, effective_date)
Rate Resolution Logic:
- When a PO is created or received in a foreign currency, the system looks up the most recent
exchange_rateentry whereeffective_date <= PO date. - If no rate exists for the currency pair, the system blocks the PO with an error: “No exchange rate found for [CURRENCY] as of [DATE]. Enter a rate in Setup > Currencies.”
- Exchange rate changes do NOT retroactively affect existing POs. The rate is captured and stored on the PO at creation time.
5.3.3 Currency Use Cases
| Use Case | Currency Behavior | Module Reference |
|---|---|---|
| Vendor Purchase Orders | PO can be denominated in vendor’s currency; line totals display in both vendor currency and base currency | Module 4, Section 4.3 |
| PO Receiving / Landed Cost | Received goods are costed in base currency using the exchange rate captured at PO creation | Module 4, Section 4.4 |
| Sales Transactions | Always in base currency (USD). No foreign currency tender support. | Module 1, Section 1.3 |
| Financial Reports | Always in base currency | Module 5, Section 5.12 |
| Customer Accounts / Credit | Always in base currency | Module 2 |
5.3.4 Currency Display Format Examples
| Currency | Format | Example |
|---|---|---|
| USD (default) | $1,234.56 | Symbol before, comma thousands, period decimal |
| EUR | 1.234,56 € | Symbol after, period thousands, comma decimal |
| GBP | £1,234.56 | Symbol before, comma thousands, period decimal |
| JPY | ¥1,234 | Symbol before, comma thousands, zero decimal places |
5.4 Locations
Scope: Defining the physical topology of the tenant’s retail operation – store locations and warehouse facilities. Locations are the organizational unit for inventory, staffing, registers, and reporting. Every inventory balance, every register, every user assignment, and every transaction is scoped to a location.
5.4.1 Location Types
Two location types are supported. The type determines which operational capabilities are available at the location.
| Type | Code | Description | Capabilities |
|---|---|---|---|
| Store | STORE | Retail store location; customer-facing with POS registers and staff | Sales, returns, exchanges, customer service, inventory counts, receiving, transfers (send and receive) |
| Warehouse | WAREHOUSE | Distribution or HQ facility; receives vendor shipments, distributes to stores | Receiving, transfers (send only to stores), inventory counts, purchase order management. No customer-facing POS. |
Business Rules:
- A warehouse location cannot have registers assigned (no POS capability).
- A warehouse location does not receive inbound transfers from stores. Transfer direction is one-way: Warehouse -> Stores, and bidirectional between Stores (Store <-> Store).
- A tenant must have at least one
STORElocation to process sales. - The
HQwarehouse receives vendor shipments and distributes stock to retail stores.
Cross-Reference: The HQ-as-warehouse pattern is documented in the SalesSight inventory analysis methodology. HQ is the distribution hub, not a retail location. Online orders displayed as “HQ sales” are fulfilled from physical stores.
5.4.2 Location Data Model
locations Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
code | String(10) | Yes | – | Unique short code within tenant (e.g., GM, HM, LM, NM, HQ) |
name | String(100) | Yes | – | Display name (e.g., “Garden Mall”, “Heritage Mall”, “Headquarters”) |
type | Enum | Yes | – | STORE or WAREHOUSE |
address_line_1 | String(200) | Yes | – | Street address |
address_line_2 | String(200) | No | null | Suite, unit, floor |
city | String(100) | Yes | – | City |
state | String(50) | Yes | – | State or province |
zip | String(20) | Yes | – | Postal / ZIP code |
country | String(2) | Yes | US | ISO 3166-1 alpha-2 country code |
phone | String(20) | No | null | Location phone number |
email | String(200) | No | null | Location email address |
timezone | String(50) | Yes | Tenant default | IANA timezone identifier; overrides tenant default timezone |
tax_jurisdiction_id | UUID | Yes | – | FK to tax_jurisdictions table; defines the compound tax jurisdiction for this location. See Section 5.9 for tax jurisdiction and rate configuration. |
is_active | Boolean | Yes | true | Whether this location is operational |
is_franchise | Boolean | Yes | false | Indicates whether this location operates as a franchise (true) or is company-owned (false). Franchise locations may have different operational rules, reporting requirements, and fee structures. |
sort_order | Integer | Yes | 0 | Display ordering in dropdowns and reports |
created_at | DateTime | Yes | Auto | Record creation timestamp |
updated_at | DateTime | Yes | Auto | Last modification timestamp |
Unique constraint: (tenant_id, code)
Cross-Reference: See Module 5, Section 5.9 for tax jurisdiction and compound rate configuration. Each location references a
tax_jurisdictionsrecord viatax_jurisdiction_id, which defines the compound tax rates (State + County + City) applied at that location. See Section 5.9.1 for thetax_jurisdictionsandtax_ratestables.
5.5 Users & Roles
Scope: Defining user profiles, authentication credentials, role-based access control, location assignments, and the feature toggle matrix that governs what each role can and cannot do within the system. Users are the human operators of the POS platform; roles determine their permissions. Every action in the system is attributable to a specific user, and every capability is gated by that user’s assigned role and feature toggles.
5.5.1 User Profile
Each user represents a staff member, manager, or administrator who interacts with the POS system. Users authenticate via email/password for the admin portal and via PIN for the POS terminal.
users Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
email | String(200) | Yes | – | Login email; unique per tenant |
password_hash | String(255) | Yes | – | Bcrypt-hashed password for admin portal login |
display_name | String(100) | Yes | – | Full name displayed in UI, receipts, and reports |
pin | String(6) | Yes | – | 4-6 digit numeric PIN for POS terminal login; stored as bcrypt hash |
role_id | UUID | Yes | – | FK to roles table |
employee_id | String(20) | No | null | Tenant-assigned employee identifier (e.g., badge number, payroll ID) |
commission_rate_percent | Decimal(5,2) | No | null | Default commission rate for this user (e.g., 5.00 for 5%). Null = no commission. |
default_register_id | UUID | No | null | FK to registers table; preferred register for auto-assignment at login |
hire_date | Date | No | null | Employee hire date for reporting and tenure tracking |
phone | String(20) | No | null | Contact phone number |
avatar_url | String(500) | No | null | URL to user avatar image for POS display |
is_active | Boolean | Yes | true | Whether user can log in; deactivated users cannot authenticate |
last_login_at | DateTime | No | null | Timestamp of most recent successful login |
failed_login_count | Integer | Yes | 0 | Consecutive failed login attempts; resets on successful login |
locked_until | DateTime | No | null | If set, user is locked out until this timestamp |
created_at | DateTime | Yes | Auto | Record creation timestamp |
updated_at | DateTime | Yes | Auto | Last modification timestamp |
Unique constraint: (tenant_id, email), (tenant_id, pin)
Business Rules:
- PIN must be unique within a tenant. Two users at the same tenant cannot share a PIN.
- Deactivating a user (
is_active = false) immediately invalidates all active sessions. The user cannot log in until reactivated. - Deleting a user is not supported; users are deactivated. All historical transactions, audit entries, and reports retain the user reference.
commission_rate_percentis the user’s default rate. Module 1, Section 1.14 details how commission is calculated per transaction and how the rate can be overridden per sale.
Cross-Reference: See Module 1, Section 1.14 for commission calculation logic including split commissions and override rates.
5.5.2 User-Location Assignment
Users can be assigned to one or more locations. The primary location determines which location the POS terminal defaults to at login. Multi-location users (e.g., managers overseeing two stores) can switch locations within the UI.
user_locations Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
user_id | UUID | Yes | – | FK to users table |
location_id | UUID | Yes | – | FK to locations table |
is_primary | Boolean | Yes | false | Whether this is the user’s primary/default location (exactly one per user) |
created_at | DateTime | Yes | Auto | Record creation timestamp |
Unique constraint: (user_id, location_id)
Business Rules:
- Every active user must have at least one location assignment.
- Exactly one location assignment per user must have
is_primary = true. - Location assignments are informational and used for default location selection, reporting filters, and organizational grouping. They do not restrict transaction processing — any user can process transactions at any location within their tenant.
- Removing a user’s last location assignment automatically deactivates the user.
5.5.3 Predefined Roles
The system ships with five predefined roles. These cover the standard organizational structure of a multi-store retail operation. Roles are tenant-scoped: each tenant gets their own copy of the predefined roles at onboarding, which they can then customize via the feature toggle matrix.
roles Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
code | String(20) | Yes | – | Unique role code within tenant |
name | String(50) | Yes | – | Display name |
description | String(500) | No | null | Role description |
is_system | Boolean | Yes | true | System roles cannot be deleted (but can be customized via feature toggles) |
created_at | DateTime | Yes | Auto | Record creation timestamp |
updated_at | DateTime | Yes | Auto | Last modification timestamp |
Unique constraint: (tenant_id, code)
Predefined Role Definitions
| Role | Code | Description | Typical Users |
|---|---|---|---|
| Staff | STAFF | POS operator; processes sales, returns, exchanges, and basic inventory tasks | Sales associates, cashiers |
| Manager | MANAGER | Store manager; approves adjustments, refunds, and price overrides; accesses location-level reports | Store managers, assistant managers |
| Admin | ADMIN | System administrator; configures all Module 5 settings, manages users and locations | IT staff, operations director |
| Buyer | BUYER | Procurement specialist; creates and manages purchase orders, vendor relationships | Purchasing agents, buyers |
| Owner | OWNER | Full access; all permissions including financial reports, audit logs, and read-only access to all configuration | Business owner, CEO |
5.5.4 Feature Toggle Matrix
The feature toggle matrix provides granular control over which capabilities each role has access to. Toggles are configured at the tenant level per role – changing a toggle affects all users with that role across the tenant.
role_feature_toggles Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
role_id | UUID | Yes | – | FK to roles table |
feature_code | String(50) | Yes | – | Feature identifier (e.g., process_sale, void_transaction) |
is_enabled | Boolean | Yes | – | Whether this feature is enabled for this role |
updated_by | UUID | No | null | FK to users table; last user who modified this toggle |
updated_at | DateTime | Yes | Auto | Timestamp of last modification |
Unique constraint: (tenant_id, role_id, feature_code)
Default Feature Toggle Configuration
The following matrix defines the default state of each feature toggle per role. Administrators can override any toggle. A checkmark indicates enabled by default; an X indicates disabled by default. All toggles are mutable.
| Feature Code | Description | Staff | Manager | Admin | Buyer | Owner |
|---|---|---|---|---|---|---|
process_sale | Create and complete a sale transaction | Y | Y | Y | N | Y |
process_return | Process merchandise returns with refund | Y | Y | Y | N | Y |
process_exchange | Process merchandise exchanges | Y | Y | Y | N | Y |
apply_discount | Apply line-item or cart-level discounts | Y | Y | Y | N | Y |
void_transaction | Void a completed same-day transaction | N | Y | Y | N | Y |
process_layaway | Create and manage layaway transactions | Y | Y | Y | N | Y |
inventory_count | Participate in physical inventory counts | Y | Y | Y | N | N |
inventory_adjust | Create manual inventory adjustments | N | Y | Y | N | N |
create_transfer | Initiate inter-store inventory transfers | N | Y | Y | N | Y |
approve_transfer | Approve outbound transfer requests | N | Y | Y | N | Y |
create_po | Create vendor purchase orders | N | N | Y | Y | Y |
approve_po | Approve purchase orders for submission | N | Y | Y | N | Y |
receive_shipment | Process inbound receiving (PO or transfer) | Y | Y | Y | Y | N |
price_change | Modify product pricing in the catalog | N | Y | Y | N | Y |
view_reports | Access reporting dashboards and exports | N | Y | Y | Y | Y |
manage_users | Create, edit, deactivate user accounts | N | N | Y | N | Y |
manage_settings | Modify Module 5 configuration settings | N | N | Y | N | Y |
view_cost_data | View product cost, margin, and vendor pricing | N | Y | Y | Y | Y |
manage_customers | Create and edit customer profiles | Y | Y | Y | N | Y |
view_audit_log | Access system audit trail | N | N | Y | N | Y |
manage_gift_cards | Issue and adjust gift card balances | N | Y | Y | N | Y |
override_price | Override selling price at POS beyond discount limits | N | Y | Y | N | Y |
cash_drawer_operations | Open cash drawer, perform cash drops, reconcile drawer | Y | Y | Y | N | Y |
end_of_day | Execute end-of-day close procedures | N | Y | Y | N | Y |
Business Rules:
- Feature toggles are evaluated at runtime. Changing a toggle takes effect immediately for all active sessions of users with that role.
- The
OWNERrole cannot havemanage_settingsormanage_userstoggled off – these are locked totruefor the OWNER role to prevent lockout. - Custom roles can be created by duplicating an existing role’s toggle configuration and modifying it. Custom roles have
is_system = false.
5.5.5 User Authentication Flow
sequenceDiagram
autonumber
participant U as User
participant POS as POS Terminal
participant API as Backend
participant DB as DB
Note over U, DB: POS Terminal Login (PIN-based)
U->>POS: Enter 4-6 Digit PIN
POS->>API: POST /auth/pin-login {pin, register_id}
API->>DB: Lookup user by PIN hash + tenant_id
alt User Found & Active
API->>DB: Check failed_login_count < max threshold
alt Not Locked
API->>DB: Reset failed_login_count = 0
API->>DB: Update last_login_at
API->>DB: Load role + feature toggles
API-->>POS: Auth Token + User Profile + Permissions
POS->>POS: Render POS UI with role-appropriate menus
else Account Locked
API-->>POS: "Account locked. Try again in X minutes."
end
else User Not Found or Inactive
API->>DB: Increment failed_login_count (by register/IP)
API-->>POS: "Invalid PIN"
end
Cross-Reference: See Module 5, Section 5.6 for clock-in/clock-out time tracking.
5.6 Time Tracking (Clock-In / Clock-Out)
Scope: Recording staff clock-in and clock-out times for basic time tracking and payroll reporting. The system provides a simple punch-clock model — staff clock in via the POS terminal using their PIN, and clock out at the end of their work period. This section does not implement shift scheduling, shift types, or workforce management.
5.6.1 Clock-In / Clock-Out
clock_records Table:
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | FK to tenants table; owning tenant |
user_id | UUID | Yes | – | FK to users table; employee who clocked in/out |
location_id | UUID | Yes | – | FK to locations table; location where clock event occurred |
clock_in | DateTime | Yes | – | Timestamp when user clocked in |
clock_out | DateTime | No | null | Timestamp when user clocked out (null = currently clocked in) |
notes | Text | No | null | Optional notes (e.g., reason for late clock-out, manager override note) |
created_at | DateTime | Yes | Auto | Record creation timestamp |
Business Rules:
- A user cannot clock in if they are already clocked in at any location (must clock out first).
- Clock-out is required before end-of-day close procedures can complete at the location.
- If a user forgets to clock out, a manager can manually enter the clock-out time with an audit note in the
notesfield. - Clock-in records are retained indefinitely for payroll and audit purposes.
- Maximum clock-in duration: 16 hours. If no clock-out is recorded within 16 hours, the system sends an alert to the location manager.
Cross-Reference: See Module 1, Section 1.8 for end-of-day cash drawer procedures that typically coincide with clock-out.
5.7 Registers & Terminals
Scope: Defining the register registry, device pairing, register profiles that control available functionality, and peripheral device assignments. A register is the logical unit representing a point-of-sale station at a location. Each register is paired with one or more physical devices and linked to peripherals (printers, scanners, payment terminals, cash drawers). Register profiles determine which POS functions are available on each terminal type.
5.7.1 Register Registry
Each location maintains a numbered set of registers. Registers are logical entities that persist across hardware replacements – when a device is swapped, the register retains its identity, transaction history, and peripheral assignments.
registers Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
location_id | UUID | Yes | – | FK to locations table |
register_number | String(20) | Yes | – | Unique register identifier within location (e.g., REG-001, REG-002) |
name | String(100) | No | null | Friendly name (e.g., “Main Counter”, “Back Register”, “Floor Mobile 1”) |
profile_id | UUID | Yes | – | FK to register_profiles table; determines available functions |
status | Enum | Yes | ACTIVE | ACTIVE, MAINTENANCE, RETIRED |
ip_address | String(45) | No | null | Network IP address of the physical device paired to this register. Supports IPv4 and IPv6. |
notes | Text | No | null | Administrative notes (e.g., “New iPad deployed 2026-01-15”) |
created_at | DateTime | Yes | Auto | Record creation timestamp |
updated_at | DateTime | Yes | Auto | Last modification timestamp |
Unique constraint: (tenant_id, location_id, register_number)
Business Rules:
- A register in
MAINTENANCEstatus cannot accept new transactions. Active sessions are preserved but no new sales can be initiated. - A register in
RETIREDstatus is permanently decommissioned. It cannot be reactivated. Its transaction history is preserved. - Registers cannot be deleted; only retired. This ensures audit trail integrity.
- Warehouse-type locations cannot have registers assigned.
- A register’s network IP address (
ip_address) can be modified a maximum of 2 times within any rolling 365-day period. IP changes are automatically tracked in theregister_ip_changesaudit table. Before updating the IP address, the system queries:SELECT COUNT(*) FROM register_ip_changes WHERE register_id = :id AND changed_at >= NOW() - INTERVAL '365 days'. IfCOUNT >= 2, the update is rejected with error:[ERR-5071] IP address change limit reached. A register's IP address can only be changed 2 times per year. Contact the system owner for an override.
register_ip_changes Table:
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | FK to tenants table; owning tenant |
register_id | UUID | Yes | – | FK to registers table |
old_ip | String(45) | No | null | Previous IP address (null for first assignment) |
new_ip | String(45) | Yes | – | New IP address being assigned |
changed_by | UUID | Yes | – | FK to users table; user who made the change |
changed_at | DateTime | Yes | Auto | Timestamp of the IP change |
5.7.2 Register State Machine
stateDiagram-v2
[*] --> ACTIVE: Register Created
ACTIVE --> MAINTENANCE: take_offline
MAINTENANCE --> ACTIVE: bring_online
ACTIVE --> RETIRED: decommission
MAINTENANCE --> RETIRED: decommission
RETIRED --> [*]
note right of ACTIVE
Accepting transactions
Device paired and online
Peripherals connected
end note
note right of MAINTENANCE
Temporarily offline
No new transactions
Active sessions preserved
Hardware swap / repair
end note
note right of RETIRED
Permanently decommissioned
Cannot reactivate
History preserved
end note
State Transition Rules:
| Transition | From | To | Trigger | Authorization | Side Effects |
|---|---|---|---|---|---|
take_offline | ACTIVE | MAINTENANCE | Manual (admin/manager) | ADMIN or MANAGER role | Active sessions warned; no new sales |
bring_online | MAINTENANCE | ACTIVE | Manual (admin/manager) | ADMIN or MANAGER role | Register available for transactions |
decommission | ACTIVE or MAINTENANCE | RETIRED | Manual (owner only) | OWNER role only | Register permanently disabled; device pairing cleared; requires type-to-confirm (see below) |
Register Retirement Safety: Register retirement (decommission) is restricted to the OWNER role only. When the owner initiates retirement, the system displays a confirmation dialog with the following warning:
‘This action permanently retires this register. Retired registers cannot be reactivated. All transaction history will be preserved but no new transactions can be processed. This action cannot be undone or reverted.’
The owner must type the word RETIRE (case-sensitive, exact match) in a confirmation text field before the system proceeds. If the typed text does not match exactly, the action is blocked with error: [ERR-5072] Confirmation text does not match. Type RETIRE to confirm.
5.7.3 Device Pairing
Each register is associated with one or more physical devices. Devices are the hardware (iPad, PC terminal, mobile phone) on which the POS application runs. A register can have multiple paired devices (e.g., a backup iPad) but only one may be the active/primary device at any time.
devices Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
register_id | UUID | Yes | – | FK to registers table |
hardware_id | String(100) | Yes | – | Unique device identifier (serial number, UDID, or system-generated fingerprint) |
device_type | Enum | Yes | – | IPAD, PC_TERMINAL, MOBILE, ANDROID_TABLET |
device_name | String(100) | No | null | Friendly name (e.g., “iPad Pro 12.9 - Main Counter”) |
os_version | String(50) | No | null | Operating system version (e.g., “iPadOS 17.4”, “Windows 11”) |
app_version | String(20) | No | null | POS application version installed (e.g., “2.3.1”) |
is_primary | Boolean | Yes | false | Whether this is the active device for the register (exactly one per register) |
last_seen_at | DateTime | No | null | Last successful heartbeat or API call from this device |
last_sync_at | DateTime | No | null | Last successful data synchronization timestamp |
paired_at | DateTime | Yes | Auto | When this device was first paired to the register |
paired_by | UUID | Yes | – | FK to users table; admin who paired the device |
Unique constraint: (tenant_id, hardware_id)
Business Rules:
- A
hardware_idcan only be paired to one register at a time across the entire tenant. Pairing to a new register automatically unpairs from the previous register. - Exactly one device per register must be
is_primary = true. When a new device is set as primary, the previous primary is automatically set tois_primary = false. - A device that has not sent a heartbeat in 5 minutes is flagged as “Offline” in the admin dashboard. After 24 hours without contact, the device status is escalated to “Disconnected” with an alert to the admin.
app_versionis reported by the device at each heartbeat. The admin dashboard highlights devices running outdated versions.
5.7.4 Register Profiles
Register profiles define which POS functions are available on a terminal. Two profiles are provided by default; tenants cannot create custom profiles (this prevents an explosion of untested UI configurations).
register_profiles Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
code | String(20) | Yes | – | Unique profile code within tenant |
name | String(50) | Yes | – | Display name |
description | String(500) | No | null | Profile description |
allowed_functions | JSONB | Yes | – | Array of function codes available on this profile |
is_system | Boolean | Yes | true | System profiles cannot be deleted |
created_at | DateTime | Yes | Auto | Record creation timestamp |
updated_at | DateTime | Yes | Auto | Last modification timestamp |
Unique constraint: (tenant_id, code)
Default Profile Definitions
| Profile | Code | Description | Available Functions |
|---|---|---|---|
| Full POS | FULL_POS | Standard counter terminal with complete POS capability | sale, return, exchange, layaway, hold, void, inventory_lookup, price_check, cash_drawer, customer_management, gift_card, reports, end_of_day, park_sale, special_order |
| Mobile Checkout | MOBILE | Handheld device for floor-based sales assistance | sale, price_check, inventory_lookup, customer_lookup, park_sale |
Function availability comparison:
| Function | Full POS | Mobile |
|---|---|---|
sale | Y | Y |
return | Y | N |
exchange | Y | N |
layaway | Y | N |
hold | Y | N |
void | Y | N |
inventory_lookup | Y | Y |
price_check | Y | Y |
cash_drawer | Y | N |
customer_management | Y | N |
customer_lookup | Y | Y |
gift_card | Y | N |
reports | Y | N |
end_of_day | Y | N |
park_sale | Y | Y |
special_order | Y | N |
Business Rules:
- The register profile controls which menu items and action buttons are rendered on the POS UI. Functions not in the profile’s
allowed_functionsarray are hidden from the interface entirely. - User role permissions (Section 5.5.4) are enforced IN ADDITION TO profile restrictions. A function must be allowed by BOTH the register profile AND the user’s role feature toggles. For example, a Staff user on a Full POS terminal cannot void a transaction because their role toggle
void_transaction = false, even though the profile allows thevoidfunction.
5.7.5 Peripheral Assignments
Each register has linked peripheral devices that provide physical I/O capabilities. Peripherals are assigned to registers via a junction table that references the device registry from the appropriate configuration section.
register_peripherals Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key, system-generated |
tenant_id | UUID | Yes | – | Owning tenant |
register_id | UUID | Yes | – | FK to registers table |
peripheral_type | Enum | Yes | – | Type of peripheral (see enumeration below) |
peripheral_ref_id | UUID | No | null | FK to the peripheral’s registry table (e.g., printers.id, payment_terminals.id) |
connection_type | Enum | Yes | – | USB, BLUETOOTH, NETWORK, BUILT_IN |
is_active | Boolean | Yes | true | Whether this peripheral is currently connected and operational |
last_status_check | DateTime | No | null | Last successful peripheral status check |
created_at | DateTime | Yes | Auto | Record creation timestamp |
Unique constraint: (register_id, peripheral_type) – one peripheral per type per register
Peripheral Type Enumeration
| Peripheral Type | Code | Required (Full POS) | Required (Mobile) | Source Configuration |
|---|---|---|---|---|
| Receipt Printer | RECEIPT_PRINTER | Yes | No | Section 5.8 (Printer Configuration) |
| Label Printer | LABEL_PRINTER | No | No | Section 5.8 (Printer Configuration) |
| Barcode Scanner | BARCODE_SCANNER | Yes | Yes | Direct pairing (USB/Bluetooth) |
| Payment Terminal | PAYMENT_TERMINAL | Yes | Yes | Section 5.11 (Payment Processing) |
| Cash Drawer | CASH_DRAWER | Yes | No | Direct pairing (connected via receipt printer kick cable) |
| Customer Display | CUSTOMER_DISPLAY | No | No | Direct pairing (secondary screen, pole display) |
| RFID Reader | RFID_READER | No | No | Section 5.16 (RFID Configuration) |
Business Rules:
- A register with the
FULL_POSprofile must have at minimum: receipt printer, barcode scanner, payment terminal, and cash drawer assigned and active before it can process transactions. - A register with the
MOBILEprofile must have at minimum: barcode scanner and payment terminal assigned. - Missing required peripherals are flagged on the admin dashboard. The POS terminal displays a warning on login: “Register [REG-001] is missing required peripheral: [Cash Drawer]. Some functions may be unavailable.”
- Cash drawers are typically connected to the receipt printer via a kick cable (RJ12 connector). The cash drawer opens when the receipt printer sends a drawer kick signal. For this reason, the cash drawer’s operational status is dependent on the receipt printer’s status.
5.7.6 Register-Peripheral Entity Relationship
erDiagram
REGISTERS ||--o{ DEVICES : "paired with"
REGISTERS ||--|| REGISTER_PROFILES : "uses profile"
REGISTERS ||--o{ REGISTER_PERIPHERALS : "has peripherals"
LOCATIONS ||--o{ REGISTERS : "contains"
REGISTERS {
UUID id PK
UUID tenant_id FK
UUID location_id FK
String register_number
UUID profile_id FK
Enum status
}
DEVICES {
UUID id PK
UUID register_id FK
String hardware_id
Enum device_type
Boolean is_primary
DateTime last_seen_at
}
REGISTER_PROFILES {
UUID id PK
String code
String name
JSONB allowed_functions
}
REGISTER_PERIPHERALS {
UUID id PK
UUID register_id FK
Enum peripheral_type
UUID peripheral_ref_id
Enum connection_type
Boolean is_active
}
LOCATIONS {
UUID id PK
String code
String name
Enum type
}
Cross-Reference: See Module 1 for POS transaction flow and how register context (profile, peripherals) affects the sales workflow. See Module 5, Section 5.8 for the printer registry that
peripheral_ref_idreferences for receipt and label printers. See Module 5, Section 5.11 for the payment terminal configuration thatperipheral_ref_idreferences for payment devices.
5.8 Printers & Peripherals
Scope: Central registry of all printers across all tenant locations, linking printers to registers, and managing network printer discovery. This section covers receipt printers, label printers, and the register-to-printer assignment model.
Cross-Reference: See Module 5, Section 5.7 for register profile definitions and peripheral assignments. See Module 5, Section 5.14 for receipt layout and content configuration.
5.8.1 Printer Registry
Every physical printer in the organization is registered in a central table. Printers are scoped to a specific location and classified by type and connection method.
Printer Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
location_id | UUID | Yes | FK to locations table – physical location where the printer resides |
name | String(100) | Yes | Human-readable name (e.g., “Main Counter Printer”, “Back Office Label Printer”) |
type | Enum | Yes | RECEIPT, LABEL |
connection_type | Enum | Yes | USB, NETWORK_IP, BLUETOOTH |
connection_address | String(255) | Yes | Connection target: IP:port for NETWORK_IP (e.g., “192.168.1.50:9100”), device path for USB (e.g., “/dev/usb/lp0”), MAC address for BLUETOOTH |
model | String(100) | No | Manufacturer model identifier (e.g., “Epson TM-T88VI”, “Zebra ZD421”, “Star TSP143IV”) |
paper_width | Enum | Yes | Receipt: 58MM, 80MM. Label: 25x50MM, 50x25MM, 50x75MM, CUSTOM |
is_shared | Boolean | Yes | Whether multiple registers can use this printer simultaneously (default: false) |
is_active | Boolean | Yes | Soft-delete flag (default: true) |
last_health_check | DateTime | No | Timestamp of the most recent successful health check ping |
last_health_status | Enum | No | ONLINE, OFFLINE, ERROR, UNKNOWN |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
5.8.2 Receipt Printers
Receipt printers produce transaction receipts, X/Z-reports, and end-of-day summaries. Every register with a Full POS profile (see Section 5.7) requires exactly one primary receipt printer.
Receipt Printer Specifications:
| Attribute | Options | Notes |
|---|---|---|
| Paper Width | 58mm (compact), 80mm (standard) | Configured per printer; determines receipt layout column width |
| Connection | USB (direct), Network IP (shared) | USB printers are 1:1 with a register; Network IP printers can be shared |
| Print Speed | Varies by model | Minimum recommended: 200mm/sec for high-volume registers |
| Auto-Cutter | Required for all receipt printers | Full or partial cut supported |
| Cash Drawer Kick | Supported via printer relay | Printer sends electrical pulse to open drawer on receipt print |
Business Rules:
- Each register in a Full POS profile must be linked to exactly one primary receipt printer. Registers with a Mobile POS or Inventory-Only profile do not require a receipt printer.
- A single receipt printer may serve multiple registers only if
is_shared = trueandconnection_type = NETWORK_IP. USB printers cannot be shared. - When a receipt printer is marked
is_active = false, any register linked to it as primary receipt printer will display a configuration warning on the POS terminal dashboard.
5.8.3 Label Printers
Label printers produce barcode labels, price tags, and shelf tags. Label printers are typically shared resources used from the back office or receiving area, though they may also be linked to individual registers.
Supported Label Sizes:
| Size Code | Dimensions | Common Use |
|---|---|---|
25x50MM | 1“ x 2“ | Small barcode labels, jewelry tags |
50x25MM | 2“ x 1“ | Standard shelf tags, price labels |
50x75MM | 2“ x 3“ | Hang tags with barcode + price + description |
CUSTOM | User-defined | Tenant-configured custom dimensions |
Business Rules:
- Label printers are always shared (
is_shared = trueby default) and can be linked to multiple registers. - Label template selection is driven by the label size configured on the printer and the template definitions in Module 3, Section 3.10.
- A location may have zero or more label printers. Label printing is optional – stores without a label printer can still operate but cannot print labels locally.
Cross-Reference: See Module 3, Section 3.10 for label template definitions, barcode symbology, and print queue management.
5.8.4 Register-Printer Linking
The register_printers junction table defines which printers are available to each register and in what role.
Register-Printer Assignment Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
register_id | UUID | Yes | FK to registers table |
printer_id | UUID | Yes | FK to printers table |
printer_role | Enum | Yes | PRIMARY_RECEIPT, LABEL, SECONDARY_RECEIPT |
is_default | Boolean | Yes | Whether this is the default printer for the given role (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
Printer Role Definitions:
| Role | Required | Max Per Register | Description |
|---|---|---|---|
PRIMARY_RECEIPT | Yes (Full POS) | 1 | Main receipt printer for transactions, X/Z-reports |
LABEL | No | 1 | Label printer for barcode/price tag printing |
SECONDARY_RECEIPT | No | 1 | Backup receipt printer; used if primary is offline |
Uniqueness Constraint: A register may have at most one printer per role. The composite key (register_id, printer_role) is unique.
5.8.5 Network Printer Discovery
Administrators can scan the local network subnet for printers to streamline the registration process.
Discovery Flow:
sequenceDiagram
autonumber
participant A as Admin
participant UI as Admin Portal
participant API as Backend
participant NET as Local Network
A->>UI: Click "Discover Printers"
UI->>API: POST /printers/discover
API->>NET: Scan subnet for devices on port 9100 (RAW), 631 (IPP)
NET-->>API: Respond with discovered IPs and device info
API-->>UI: Return discovered printer list
Note right of UI: Display: IP, hostname, model (if available), port
A->>UI: Select discovered printer
A->>UI: Assign name, type (RECEIPT/LABEL), paper width
UI->>API: POST /printers
API-->>UI: Printer registered successfully
Discovery Rules:
- Scan is limited to the local subnet of the location’s network.
- Discovery returns IP address, hostname (if resolvable), and model string (if the printer supports SNMP or IPP device identification).
- Discovered printers are presented as candidates – the administrator must explicitly add them to the registry with a name and type assignment.
- Discovery does not modify any existing printer records.
5.8.6 Printer Health Monitoring
The system performs periodic health checks on all active network printers.
| Setting | Value | Description |
|---|---|---|
| Health check interval | Every 5 minutes | Background ping to NETWORK_IP printers only |
| Offline threshold | 3 consecutive failures | Printer status changes to OFFLINE after 3 failed pings |
| Alert trigger | On status change to OFFLINE | Dashboard notification sent to location manager |
| USB printers | Not health-checked | USB status determined at print time |
Business Rules:
- Health checks apply only to printers with
connection_type = NETWORK_IP. - When a primary receipt printer goes offline, the register automatically attempts to use the secondary receipt printer (if configured).
- Printer health status is visible on the Admin Portal dashboard per location.
5.9 Tax Configuration
Scope: Location-level compound tax configuration using a 3-level jurisdiction model (State / County / City) with support for product-level and customer-level exemptions. Each location references a tax jurisdiction; all active rates within that jurisdiction are summed at time of sale to produce the effective compound rate.
Cross-Reference: See Module 1, Section 1.17 for the tax calculation engine and line-item tax computation. See Module 2 for customer tax exemption fields. See Module 3, Section 3.1 for product-level tax exemption.
5.9.1 Tax Jurisdiction and Rate Setup
Tax is modeled as a 3-level compound system: State, County, and City. Each location references a tax jurisdiction, and each jurisdiction defines up to three rate levels that are summed at time of sale to produce the effective compound tax rate.
tax_jurisdictions Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key |
tenant_id | UUID | Yes | – | FK to tenants table |
code | String(20) | Yes | – | Unique jurisdiction code (e.g., “VA-NFK”, “VA-VB”, “CA-LA”) |
name | String(100) | Yes | – | Human-readable name (e.g., “Norfolk, Virginia”) |
state_name | String(50) | Yes | – | State or province name |
description | String(500) | No | null | Additional notes |
is_active | Boolean | Yes | true | Whether available for assignment |
created_at | DateTime | Yes | Auto | Record creation timestamp |
updated_at | DateTime | Yes | Auto | Last modification timestamp |
Unique constraint: (tenant_id, code)
tax_rates Table
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | Auto | Primary key |
tenant_id | UUID | Yes | – | FK to tenants table |
jurisdiction_id | UUID | Yes | – | FK to tax_jurisdictions table |
level | Enum | Yes | – | Tax level: STATE, COUNTY, CITY |
rate_name | String(100) | Yes | – | Display name (e.g., “Virginia State Tax”, “Norfolk City Tax”) |
rate_percent | Decimal(5,3) | Yes | – | Tax rate as percentage (e.g., 4.300 for 4.3%) |
effective_date | Date | Yes | – | Date this rate becomes active |
end_date | Date | No | null | Date this rate expires (null = no expiry) |
created_by | UUID | Yes | – | FK to users table |
is_active | Boolean | Yes | true | Whether currently in effect |
notes | String(500) | No | null | Reason for rate or change |
created_at | DateTime | Yes | Auto | Record creation timestamp |
Unique constraint: (jurisdiction_id, level, effective_date)
Business Rules:
- A jurisdiction may have up to 3 active tax rate levels (
STATE,COUNTY,CITY). Not all levels are required. - At time of sale, all active rates for the location’s jurisdiction are summed to produce the effective compound tax rate.
- When a new rate’s
effective_datearrives, the system deactivates any existing rate at the same level without anend_date. - Future rates can be scheduled by setting
effective_datein the future. A background job activates the rate at midnight on the effective date. - Rate changes never modify historical records. All past rates are preserved for audit and historical transaction recalculation if needed.
- The system does not support tax-inclusive pricing. All product prices are tax-exclusive; tax is computed and displayed separately.
- Example: Norfolk, VA = State 4.3% + Regional 0.7% + City 1.0% = 6.0% compound rate.
5.9.2 Tax Calculation Priority
Tax determination follows a strict priority order. The first matching rule wins:
flowchart TD
A[Line Item Added to Cart] --> B{Product tax_exempt = true?}
B -->|Yes| C[Tax Amount = $0.00]
B -->|No| D{Customer attached?}
D -->|No| G[Apply Location Jurisdiction Compound Rate]
D -->|Yes| E{Customer tax_exempt = true AND certificate valid?}
E -->|Yes| F[Tax Amount = $0.00]
E -->|No| G
G --> H[Sum all active rates for jurisdiction]
H --> I["Tax = line_subtotal × sum_of_rates / 100"]
Priority Order (highest first):
| Priority | Condition | Result |
|---|---|---|
| 1 | Product tax_exempt = true | No tax on this line item |
| 2 | Customer tax_exempt = true AND exemption_certificate_expiry >= today | No tax on any line item for this customer |
| 3 | Location jurisdiction compound rate | Sum all active rates for location’s jurisdiction (State + County + City). Apply sum_of_rates to taxable line subtotal. Formula: Tax = line_subtotal × sum_of_rates / 100 |
5.9.3 Tax Exemption
Product-Level Exemption:
- The
tax_exemptboolean flag on the product record (Module 3, Section 3.1) exempts individual products from tax regardless of customer or location. - Common use: Food items, certain clothing categories in jurisdictions with clothing exemptions.
Customer-Level Exemption:
- Customer records (Module 2) include three exemption fields:
| Field | Type | Description |
|---|---|---|
tax_exempt | Boolean | Whether this customer is tax-exempt (default: false) |
exemption_certificate_number | String(50) | State or federal tax exemption certificate number |
exemption_certificate_expiry | Date | Expiration date of the certificate – system checks validity at time of sale |
- When a tax-exempt customer is attached to a transaction, the system validates that the certificate has not expired. If expired, the customer is treated as taxable and the cashier sees a warning: “Tax exemption certificate expired – tax will be applied.”
- Exemption applies to all line items in the transaction (unless the product itself is tax-exempt, in which case it remains exempt regardless).
5.9.4 Tax Display and Reporting
Receipt Display:
- Tax is calculated per line item and aggregated at the transaction level.
- Receipt shows: Subtotal (pre-tax) + Tax Amount = Total.
- Compound tax rate is printed on the receipt (e.g., “Tax (6.000%): $4.50”). Optionally, the breakdown by level can be shown (e.g., State 4.3%, County 0.7%, City 1.0%).
- Tax-exempt transactions display “Tax Exempt” with the certificate number.
Tax Reporting Period:
| Setting | Options | Default | Description |
|---|---|---|---|
tax_reporting_period | MONTHLY, QUARTERLY | QUARTERLY | Determines aggregation period for tax liability reports |
- Tax liability reports aggregate taxable sales, exempt sales, and tax collected by reporting period.
- Reports are available per location and consolidated across all locations.
Cross-Reference: See Module 1, Section 1.17 for detailed tax calculation engine, rounding rules, and tax line-item storage.
5.10 Units of Measure
Scope: Predefined and tenant-customizable units of measure (UoMs) used for selling, purchasing, and inventory tracking. The UoM system supports conversion factors between related units, enabling scenarios where products are purchased in bulk units (cases, dozens) and sold in individual units (each, pair).
Cross-Reference: See Module 3, Section 3.1 for product UoM assignment fields (
selling_uom,purchasing_uom,uom_conversion_factor). See Module 4, Section 4.2 for purchase order UoM handling.
5.10.1 System-Predefined UoMs
The following UoMs are provided out-of-the-box and cannot be deleted or modified. They are available to all tenants.
| Code | Name | Category | Base Unit | Conversion to Base | Example Use |
|---|---|---|---|---|---|
EACH | Each | QUANTITY | EACH | 1 (base) | Individual garments, accessories |
PAIR | Pair | QUANTITY | EACH | 2 | Shoes, gloves, earrings, socks |
PACK | Pack | QUANTITY | EACH | Varies (set per product) | Multi-pack underwear, sock bundles |
BOX | Box | QUANTITY | EACH | Varies (set per product) | Boxed gift sets, assortments |
DOZEN | Dozen | QUANTITY | EACH | 12 | Bulk socks, accessories wholesale |
CASE | Case | QUANTITY | EACH | Varies (set per product) | Vendor case packs |
YARD | Yard | LENGTH | YARD | 1 (base) | Fabric, ribbon, trim |
METER | Meter | LENGTH | YARD | 1.0936 | Fabric (metric suppliers) |
FOOT | Foot | LENGTH | YARD | 0.3333 | Chain, cord, elastic |
KG | Kilogram | WEIGHT | KG | 1 (base) | Bulk items by weight |
LB | Pound | WEIGHT | KG | 0.4536 | Bulk items (imperial) |
OZ | Ounce | WEIGHT | KG | 0.02835 | Small items, jewelry |
5.10.2 Custom UoMs
Tenants can create additional UoMs to match their specific business needs. Custom UoMs must belong to an existing category and define a conversion factor to the category’s base unit.
UoM Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table (NULL for system-predefined UoMs) |
code | String(20) | Yes | Unique code within tenant scope (e.g., “ROLL”, “SPOOL”, “BUNDLE”, “SET”) |
name | String(50) | Yes | Display name (e.g., “Roll”, “Spool”, “Bundle”, “Set of 3”) |
category | Enum | Yes | QUANTITY, LENGTH, WEIGHT |
conversion_factor | Decimal(12,6) | Yes | Number of base units in one of this UoM (e.g., ROLL = 25 YARD, so factor = 25) |
base_uom_id | UUID | Yes | FK to uom table – the base unit this converts to (EACH, YARD, or KG) |
is_system | Boolean | Yes | true for predefined UoMs, false for tenant-created (default: false) |
is_active | Boolean | Yes | Soft-delete flag (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Custom UoM Examples:
| Code | Name | Category | Base UoM | Conversion Factor | Meaning |
|---|---|---|---|---|---|
ROLL | Roll | LENGTH | YARD | 25 | 1 Roll = 25 Yards |
SPOOL | Spool | LENGTH | YARD | 100 | 1 Spool = 100 Yards |
BUNDLE | Bundle | QUANTITY | EACH | 5 | 1 Bundle = 5 Each |
SET3 | Set of 3 | QUANTITY | EACH | 3 | 1 Set = 3 Each |
HALFYD | Half Yard | LENGTH | YARD | 0.5 | 1 Half Yard = 0.5 Yards |
5.10.3 UoM Conversion Table
For complex multi-step conversions (e.g., converting between two non-base units), the system maintains an explicit conversion table.
UoM Conversion Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
from_uom_id | UUID | Yes | FK to uom table – source unit |
to_uom_id | UUID | Yes | FK to uom table – target unit |
factor | Decimal(12,6) | Yes | Multiply source quantity by this factor to get target quantity |
tenant_id | UUID | Yes | FK to tenants table |
created_at | DateTime | Yes | Record creation timestamp |
Conversion Examples:
| From | To | Factor | Explanation |
|---|---|---|---|
| DOZEN | EACH | 12 | 1 Dozen = 12 Each |
| CASE | EACH | 24 | 1 Case = 24 Each (varies per product) |
| PAIR | EACH | 2 | 1 Pair = 2 Each |
| METER | YARD | 1.0936 | 1 Meter = 1.0936 Yards |
| LB | KG | 0.4536 | 1 Pound = 0.4536 Kilograms |
| ROLL | YARD | 25 | 1 Roll = 25 Yards |
Uniqueness Constraint: The composite key (from_uom_id, to_uom_id, tenant_id) is unique. The system auto-generates the inverse conversion (e.g., if DOZEN->EACH = 12, then EACH->DOZEN = 0.083333) so both directions are always available.
5.10.4 Product UoM Assignment
Each product specifies how it is sold and how it is purchased. The conversion factor bridges these two units for inventory tracking.
Product UoM Fields (on Product Record):
| Field | Type | Required | Description |
|---|---|---|---|
selling_uom_id | UUID | Yes | FK to uom table – unit used at POS (e.g., EACH, PAIR, YARD) |
purchasing_uom_id | UUID | Yes | FK to uom table – unit used on purchase orders (e.g., CASE, DOZEN, ROLL) |
uom_conversion_factor | Decimal(12,6) | Yes | Number of selling units per purchasing unit (e.g., 24 EACH per CASE) |
Conversion in Practice:
Purchase Order: 5 CASES of "Classic V-Neck Tee"
uom_conversion_factor = 24 (1 CASE = 24 EACH)
→ Receiving adds 5 × 24 = 120 EACH to inventory
POS Sale: Customer buys 3 EACH of "Classic V-Neck Tee"
→ Inventory decremented by 3 EACH
→ Remaining: 117 EACH (or 4.875 CASES)
Business Rules:
- The
selling_uom_iddetermines how inventory quantities are displayed at the POS and in stock reports. - The
purchasing_uom_iddetermines the unit on purchase orders and receiving documents. - When receiving a PO, the system multiplies received quantity by
uom_conversion_factorto compute the inventory increment in selling units. - If
selling_uom_idequalspurchasing_uom_id, thenuom_conversion_factormust be 1. - UoM changes on a product with existing inventory require a manager approval and trigger an inventory adjustment record.
Cross-Reference: See Module 3, Section 3.1 for full product data model. See Module 4, Section 4.2 for purchase order line-item UoM handling. See Module 4, Section 4.3 for receiving UoM conversion.
5.11 Payment Methods & Processors
Scope: Configuration of accepted payment methods per location, payment processor integrations, terminal management, and cash rounding rules. This section defines the payment method registry and processor setup – the transactional payment flow is documented in Module 1.
Cross-Reference: See Module 1, Section 1.18 for payment integration flow and split-payment logic. See Module 5, Section 5.7 for register payment terminal assignment.
5.11.1 Payment Methods
The system supports a fixed set of payment method types. Each method has inherent capabilities (processor requirement, split eligibility, offline support) that cannot be overridden.
Payment Method Registry
| Code | Name | Requires Processor | Can Split | Offline Capable | Description |
|---|---|---|---|---|---|
CASH | Cash | No | Yes | Yes | Physical currency; change calculated automatically |
CREDIT_CARD | Credit/Debit Card | Yes (external) | Yes (multi-card) | No | Chip, swipe, tap, or manual entry via payment terminal |
GIFT_CARD | Gift Card | Internal | Yes | No | Store-issued gift cards with balance tracking |
STORE_CREDIT | Store Credit | Internal | Yes | No | Credit balance on customer account (from returns, adjustments) |
LAYAWAY_PAYMENT | Layaway Payment | Via card/cash | Yes | No | Partial payment applied to layaway balance |
FINANCING | Third-Party Financing | Yes (external) | No | No | Affirm, Klarna, or similar buy-now-pay-later provider |
Payment Method Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
code | String(20) | Yes | Method code from the table above |
name | String(50) | Yes | Display name (customizable by tenant, e.g., “Visa/MC/Amex” instead of “Credit/Debit Card”) |
requires_processor | Boolean | Yes | Whether an external or internal processor is required |
can_split | Boolean | Yes | Whether this method can be combined with other methods in a single transaction |
offline_capable | Boolean | Yes | Whether this method can be used when the terminal is offline |
is_active | Boolean | Yes | Global enable/disable for the tenant (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
5.11.2 Per-Location Payment Configuration
Each payment method can be independently enabled or disabled at each location. This allows tenants to offer different payment options at different store locations (e.g., financing only at the flagship store, no gift cards at popup locations).
Location Payment Method Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
location_id | UUID | Yes | FK to locations table |
payment_method_id | UUID | Yes | FK to payment_methods table |
is_enabled | Boolean | Yes | Whether this payment method is accepted at this location (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Uniqueness Constraint: The composite key (location_id, payment_method_id) is unique.
Business Rules:
- A payment method must be active at the tenant level (
payment_methods.is_active = true) AND enabled at the location level (location_payment_methods.is_enabled = true) to appear as an option on the POS terminal at that location. - CASH is always enabled and cannot be disabled at any location.
- When a payment method is disabled at the tenant level, it is automatically hidden at all locations regardless of the location-level setting.
5.11.3 Payment Processor Configuration
MOVED TO MODULE 6: Payment processor data model, terminal mapping, processor type details, and business rules have been consolidated into Module 6, Section 6.8.3 (Processor Configuration).
See: Module 6, Section 6.8 for the complete payment processor integration specification including SAQ-A architecture, terminal communication, failure handling, and batch settlement.
5.11.4 Cash Rounding Rules
When the total transaction amount results in a fractional cent, rounding rules determine how the final amount is adjusted for cash payments. Card payments are always exact (no rounding applied).
Rounding Rule Options
| Rule | Code | Description | Example |
|---|---|---|---|
| Nearest Cent | NEAREST_CENT | Round to nearest $0.01 (standard – no visible rounding) | $12.347 → $12.35 |
| Nearest Nickel | NEAREST_NICKEL | Round to nearest $0.05 (cash only) | $12.32 → $12.30; $12.33 → $12.35 |
| Nearest Dime | NEAREST_DIME | Round to nearest $0.10 (cash only) | $12.34 → $12.30; $12.36 → $12.40 |
Configuration:
| Field | Type | Required | Description |
|---|---|---|---|
tenant_id | UUID | Yes | FK to tenants table |
cash_rounding_rule | Enum | Yes | NEAREST_CENT, NEAREST_NICKEL, NEAREST_DIME (default: NEAREST_CENT) |
Business Rules:
- Rounding applies ONLY to the total amount when the payment method is CASH or includes a CASH component in a split payment.
- Card payments, gift cards, and store credits are always exact – no rounding.
- The rounding adjustment (positive or negative) is recorded as a separate line on the receipt (e.g., “Cash Rounding: -$0.02”) and tracked in the
cash_rounding_amountfield on the transaction record. - Rounding adjustments are excluded from tax calculations – tax is computed on the pre-rounding subtotal.
5.12 Custom Fields
Scope: Tenant-defined custom fields that extend the standard data model for products, customers, orders, and vendors. Custom fields provide schema flexibility without database migrations, enabling each tenant to capture business-specific attributes unique to their operation.
Cross-Reference: See Module 3, Section 3.1.4 for the original product custom attribute specification. This section generalizes that pattern to all supported entity types.
5.12.1 Supported Entity Types
| Entity Type | Module | Max Fields Per Tenant | Use Cases |
|---|---|---|---|
PRODUCT | Module 3 | 50 | Care instructions, fabric composition, country of origin, certification level, custom sizing notes |
CUSTOMER | Module 2 | 50 | Preferred size, allergies/sensitivities, referral source, VIP notes, personal shopper assignment |
ORDER | Module 1 | 20 | Delivery instructions, gift wrap preference, event name, sales associate notes |
VENDOR | Module 3 | 20 | Internal account number, EDI trading partner code, preferred contact method, payment terms notes |
5.12.2 Custom Field Definition
Each custom field is defined once at the tenant level and then applied to individual entity records.
Custom Field Definition Data Model
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | – | Primary key |
tenant_id | UUID | Yes | – | FK to tenants table – owning tenant |
entity_type | Enum | Yes | – | PRODUCT, CUSTOMER, ORDER, VENDOR |
field_name | String(50) | Yes | – | Internal key in snake_case (e.g., “care_instructions”, “referral_source”). Must be unique within entity_type + tenant. |
label | String(100) | Yes | – | Human-readable display name (e.g., “Care Instructions”, “Referral Source”) |
field_type | Enum | Yes | – | TEXT, NUMBER, DATE, DROPDOWN, BOOLEAN |
is_required | Boolean | Yes | false | Whether this field must be filled when saving the entity |
default_value | String(500) | No | NULL | Default value applied when a new entity is created (must match field_type validation) |
sort_order | Integer | Yes | 0 | Display position in the entity edit form (lower = higher) |
is_active | Boolean | Yes | true | Soft-delete flag; inactive fields are hidden from forms but data is preserved |
show_on_pos | Boolean | Yes | false | Whether this field is visible on the POS terminal (useful for quick customer notes or product care info) |
validation_min | Decimal(12,4) | No | NULL | Minimum value for NUMBER fields |
validation_max | Decimal(12,4) | No | NULL | Maximum value for NUMBER fields |
validation_max_length | Integer | No | 500 | Maximum character length for TEXT fields |
created_at | DateTime | Yes | – | Record creation timestamp |
updated_at | DateTime | Yes | – | Last modification timestamp |
5.12.3 Field Type Specifications
| Field Type | Stored Column | Validation Rules | Example Value |
|---|---|---|---|
TEXT | value_text (VARCHAR 500) | Max length enforced; blank allowed unless is_required | “Dry clean only” |
NUMBER | value_number (DECIMAL 12,4) | Must be numeric; validation_min / validation_max enforced if set | 42.5000 |
DATE | value_date (DATE) | Must be a valid ISO 8601 date | “2026-03-15” |
DROPDOWN | value_text (VARCHAR 100) | Value must match one of the defined options in custom_field_options | “Cotton” |
BOOLEAN | value_boolean (BOOLEAN) | Must be true or false | true |
5.12.4 Dropdown Options
When field_type = DROPDOWN, the allowed values are stored in a separate options table.
Custom Field Options Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
field_id | UUID | Yes | FK to custom_field_definitions table |
option_value | String(100) | Yes | The selectable value (e.g., “Cotton”, “Polyester”, “Silk”) |
sort_order | Integer | Yes | Display position in the dropdown (lower = higher) |
is_active | Boolean | Yes | Soft-delete flag (default: true); inactive options are hidden from new selections but preserved on existing records |
created_at | DateTime | Yes | Record creation timestamp |
Example Dropdown Configuration:
| Field Label | Entity Type | Options |
|---|---|---|
| Material | PRODUCT | Cotton, Polyester, Silk, Wool, Linen, Blend, Leather, Synthetic |
| Referral Source | CUSTOMER | Walk-in, Website, Social Media, Friend/Family, Google, Event, Other |
| Gift Wrap Style | ORDER | None, Standard, Premium, Holiday |
| Payment Terms | VENDOR | Net 30, Net 60, Net 90, COD, Prepaid |
5.12.5 Custom Field Values
Custom field values are stored in a generic key-value table using typed columns. Only the column matching the field’s type is populated; the others remain NULL.
Custom Field Values Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
entity_type | Enum | Yes | PRODUCT, CUSTOMER, ORDER, VENDOR – matches the definition’s entity_type |
entity_id | UUID | Yes | FK to the entity record (product, customer, order, or vendor) |
field_id | UUID | Yes | FK to custom_field_definitions table |
value_text | String(500) | No | Populated when field_type = TEXT or DROPDOWN |
value_number | Decimal(12,4) | No | Populated when field_type = NUMBER |
value_date | Date | No | Populated when field_type = DATE |
value_boolean | Boolean | No | Populated when field_type = BOOLEAN |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Uniqueness Constraint: The composite key (entity_type, entity_id, field_id) is unique – each entity record has at most one value per custom field.
Indexing Strategy:
- GIN index on
(entity_type, entity_id)for fast retrieval of all custom field values for a given entity. - B-tree index on
(field_id, value_text)for custom fields markedsearchable = true(PRODUCT entity type only).
5.12.6 Business Rules
- Field Limit: Maximum 50 custom field definitions per entity type per tenant for PRODUCT and CUSTOMER. Maximum 20 per entity type per tenant for ORDER and VENDOR. Attempting to exceed the limit returns error: “Maximum custom fields reached for [entity_type]. Archive unused fields to create new ones.”
- Archival: Setting
is_active = falseon a field definition hides it from all forms and POS screens but preserves existing values. Reactivating the field restores visibility and all previously stored values. - Deletion: Custom field definitions cannot be hard-deleted. Only soft-delete via
is_active = falseis supported. This ensures data integrity and audit compliance. - POS Visibility: Fields with
show_on_pos = trueappear in a “Custom Info” panel on the POS terminal. Maximum 5 fields per entity type can haveshow_on_pos = trueto prevent POS screen clutter. - Required Field Enforcement: When
is_required = true, the entity cannot be saved without a value for this field. For PRODUCT entities, this applies when transitioning from DRAFT to ACTIVE status (drafts may have incomplete custom fields). - Dropdown Integrity: If an active option is deactivated, existing records that reference that option retain their value (displayed with a “deprecated” indicator). New records cannot select the deactivated option.
5.13 Approval Workflows
Scope: Configurable approval rules that gate sensitive business actions behind manager or administrator review. Each approvable action has its own rule defining whether approval is required, the threshold that triggers it, who can approve, and how notifications are delivered.
Cross-Reference: See Module 1 for refund and void transaction workflows. See Module 3 for price markdown workflows. See Module 4, Section 4.3 for purchase order approval. See Module 4, Section 4.7 for inventory adjustment approval.
5.13.1 Approvable Actions
The system supports the following approvable actions. Each action is identified by a unique code and linked to a specific module.
| Action Code | Module | Description | Threshold Type | Default Threshold |
|---|---|---|---|---|
PO_CREATE | Module 4 | Purchase order creation and submission | Dollar amount | $5,000 |
PO_ABOVE_THRESHOLD | Module 4 | PO exceeding the tenant’s auto-approve limit | Dollar amount | $10,000 |
INVENTORY_ADJUSTMENT | Module 4 | Manual inventory quantity or value adjustment | Unit count or dollar value | 50 units or $500 |
PRICE_MARKDOWN | Module 3 | Price reduction on one or more products | Percentage or dollar amount | 30% or $50 |
REFUND_ABOVE_THRESHOLD | Module 1 | Refund exceeding the per-transaction refund limit | Dollar amount | $200 |
VOID_TRANSACTION | Module 1 | Voiding a completed, finalized transaction | Always (no threshold) | N/A – always requires approval |
INTER_STORE_TRANSFER | Module 4 | Transfer of inventory between locations | Unit count | 100 units |
VENDOR_RMA | Module 4 | Return merchandise authorization to vendor | Dollar amount | $1,000 |
DISCOUNT_OVERRIDE | Module 1 | Discount exceeding the maximum allowed percentage | Percentage | 25% |
5.13.2 Approval Rule Configuration
Each tenant configures one rule per approvable action. Rules can be enabled or disabled independently.
Approval Rule Data Model
| Field | Type | Required | Default | Description |
|---|---|---|---|---|
id | UUID | Yes | – | Primary key |
tenant_id | UUID | Yes | – | FK to tenants table – owning tenant |
action_code | String(50) | Yes | – | Action code from the approvable actions table (unique per tenant) |
is_enabled | Boolean | Yes | true | Whether this approval rule is active |
threshold_value | Decimal(12,2) | No | NULL | Numeric threshold that triggers the approval requirement (NULL when threshold_type = ALWAYS) |
threshold_type | Enum | Yes | – | AMOUNT (dollar), UNITS (count), PERCENT (percentage), ALWAYS (no threshold – always requires approval) |
approver_role | Enum | Yes | MANAGER | Minimum role required to approve: MANAGER, ADMIN, OWNER |
notification_method | Enum | Yes | BOTH | How the approver is notified: IN_APP, EMAIL, BOTH |
escalation_timeout_hours | Integer | No | 24 | Hours before a pending request escalates to the next higher role (NULL = no escalation) |
auto_reject_on_timeout | Boolean | Yes | false | If true, requests that exceed escalation timeout without action are auto-rejected |
created_at | DateTime | Yes | – | Record creation timestamp |
updated_at | DateTime | Yes | – | Last modification timestamp |
5.13.3 Approval Request Lifecycle
When an action triggers an approval requirement, the system creates an approval request and routes it through the following state machine.
stateDiagram-v2
[*] --> CHECK_THRESHOLD : Action Initiated
CHECK_THRESHOLD --> AUTO_APPROVED : Below Threshold
CHECK_THRESHOLD --> PENDING_APPROVAL : At or Above Threshold
PENDING_APPROVAL --> APPROVED : Approver Accepts
PENDING_APPROVAL --> REJECTED : Approver Rejects
PENDING_APPROVAL --> ESCALATED : Escalation Timeout Reached
ESCALATED --> APPROVED : Higher-Role Approver Accepts
ESCALATED --> REJECTED : Higher-Role Approver Rejects
ESCALATED --> AUTO_REJECTED : Auto-Reject on Timeout (if enabled)
AUTO_APPROVED --> [*]
APPROVED --> [*]
REJECTED --> [*]
AUTO_REJECTED --> [*]
State Definitions:
| State | Description |
|---|---|
CHECK_THRESHOLD | System evaluates the action value against the rule’s threshold |
AUTO_APPROVED | Action value is below the threshold – no human approval needed; action proceeds immediately |
PENDING_APPROVAL | Waiting for a user with the required role to review and accept or reject |
APPROVED | An authorized approver accepted the request – action proceeds |
REJECTED | An authorized approver rejected the request – action is blocked and the requester is notified with the rejection reason |
ESCALATED | The escalation_timeout_hours elapsed without action; the request is re-routed to users with the next higher role |
AUTO_REJECTED | The escalation timeout elapsed AND auto_reject_on_timeout = true – request is automatically rejected |
5.13.4 Escalation Chain
When a request escalates, the system promotes the required approver role one level up.
| Original Role | Escalates To | Final Escalation |
|---|---|---|
MANAGER | ADMIN | OWNER |
ADMIN | OWNER | No further escalation – remains pending until OWNER acts or auto-reject triggers |
OWNER | N/A | Cannot escalate; remains pending or auto-rejects |
5.13.5 Notification Behavior
| Method | Behavior |
|---|---|
IN_APP | Dashboard notification badge and entry in the “Pending Approvals” queue visible to all users with the required role at the relevant location(s) |
EMAIL | Email sent to all users with the approver role at the relevant location(s). Email includes: action description, requested by, threshold value, and a direct link to approve/reject in the Admin Portal. |
BOTH | Dashboard notification AND email are sent simultaneously |
Notification Rules:
- Notifications are scoped to the location where the action originated. If the action is tenant-wide (e.g., a PO for the entire organization), notifications go to all users with the approver role across all locations.
- When a request escalates, a new notification is sent to users with the escalated role. The original notification is updated to show “Escalated.”
- Upon approval or rejection, the requester receives a notification with the decision and any rejection reason.
5.13.6 Approval Request Data Model
Approval Request Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
action_code | String(50) | Yes | Action code that triggered this request |
status | Enum | Yes | PENDING, APPROVED, REJECTED, ESCALATED, AUTO_APPROVED, AUTO_REJECTED |
requested_by | UUID | Yes | FK to users table – user who initiated the action |
requested_at | DateTime | Yes | Timestamp when the request was created |
approved_by | UUID | No | FK to users table – user who approved or rejected (NULL while pending) |
approved_at | DateTime | No | Timestamp of approval/rejection |
rejection_reason | String(500) | No | Free-text reason provided by the approver when rejecting |
reference_type | Enum | Yes | Entity type the request relates to: PO, ADJUSTMENT, TRANSACTION, TRANSFER, RMA, PRODUCT, DISCOUNT |
reference_id | UUID | Yes | FK to the specific entity record (purchase order, transaction, adjustment, etc.) |
threshold_value_at_time | Decimal(12,2) | No | The actual value that triggered the approval (e.g., PO total, refund amount) – captured at request time for audit |
location_id | UUID | No | FK to locations table – location where the action originated (NULL for tenant-wide actions) |
escalated_at | DateTime | No | Timestamp when the request was escalated (NULL if not escalated) |
created_at | DateTime | Yes | Record creation timestamp |
Business Rules:
- Approval requests are immutable once resolved (APPROVED, REJECTED, or AUTO_REJECTED). The status cannot be changed after resolution.
- A user cannot approve their own request – the
approved_byuser must be different from therequested_byuser. - When
VOID_TRANSACTIONis the action, approval is always required regardless of transaction amount (threshold_type = ALWAYS). - Approval requests older than 90 days in PENDING or ESCALATED status are automatically rejected with reason: “Request expired – no action taken within 90 days.”
- The
threshold_value_at_timefield captures the actual value at request creation, ensuring accurate audit even if the underlying rule’s threshold is later changed.
5.14 Receipt Configuration
Scope: Full customization of receipt layout, content, and formatting for both printed thermal receipts and email receipts. Receipt configuration is set at the tenant level with optional location-level overrides.
Cross-Reference: See Module 5, Section 5.8 for receipt printer hardware configuration. See Module 1 for transaction receipt generation and print trigger logic.
5.14.1 Receipt Field Registry
Each field on the receipt can be independently toggled (shown or hidden) and reordered. The following fields are available.
| Field Code | Default Show | Category | Description |
|---|---|---|---|
store_name | Yes | Header | Store or location name |
store_address | Yes | Header | Store street address, city, state, zip (from location configuration, Section 5.3) |
store_phone | Yes | Header | Store phone number (from location configuration) |
cashier_name | Yes | Transaction | Name of the staff member who processed the sale |
register_number | No | Transaction | Register identifier (e.g., “Register 3”) |
transaction_number | Yes | Transaction | Unique transaction ID (e.g., “TXN-2026-001234”) |
transaction_date | Yes | Transaction | Date and time of the transaction |
barcode | Yes | Transaction | Scannable CODE-128 barcode encoding the transaction number (for easy lookup on returns) |
item_list | Yes | Line Items | Itemized list showing: item name, SKU, quantity, unit price, line discount (if any), line total |
subtotal | Yes | Totals | Pre-tax total of all line items |
discount_total | Yes | Totals | Total discount amount applied (shown only if > $0.00) |
tax_amount | Yes | Totals | Tax line showing rate and amount (e.g., “Tax (6.000%): $4.50”) |
total | Yes | Totals | Grand total (subtotal - discounts + tax) |
payment_details | Yes | Payment | Payment method(s) used and amount per method (e.g., “Visa ****1234: $45.00, Cash: $10.00”) |
change_due | Yes | Payment | Change amount returned to customer (shown only for cash payments with overpayment) |
loyalty_points | No | Loyalty | Points earned on this transaction and current balance (shown only if loyalty module is enabled) |
customer_name | No | Customer | Customer name (shown only if a customer is attached to the transaction) |
savings_total | No | Totals | “You saved $X.XX” message showing total promotional and discount savings |
5.14.2 Layout Settings
Receipt Layout Data Model
| Setting | Type | Options | Default | Description |
|---|---|---|---|---|
paper_width | Enum | 58MM, 80MM | 80MM | Paper width – determines character-per-line limit (58mm = ~32 chars, 80mm = ~48 chars) |
font_size | Enum | SMALL, MEDIUM, LARGE | MEDIUM | Print font size – affects line density and readability |
field_order | JSON Array | Array of field_code strings | Default order from field registry | Ordered list defining top-to-bottom print sequence |
line_separator | Enum | DASH, EQUALS, BLANK, STAR | DASH | Character used to separate receipt sections |
alignment | Enum | LEFT, CENTER | CENTER | Header and footer text alignment |
print_density | Enum | LIGHT, NORMAL, DARK | NORMAL | Thermal print darkness (affects readability and paper consumption) |
Line Separator Examples:
| Option | Rendered As |
|---|---|
DASH | -------------------------------- |
EQUALS | ================================ |
BLANK | (empty line) |
STAR | ******************************** |
5.14.3 Header Configuration
The receipt header appears at the top of every printed receipt and supports up to 3 customizable text lines plus an optional logo.
| Field | Type | Max Length | Default | Description |
|---|---|---|---|---|
header_line_1 | String | 100 chars | Tenant name | Primary header text (typically the company name) |
header_line_2 | String | 100 chars | “Thank you for shopping with us!” | Secondary header text (tagline, greeting, or blank) |
header_line_3 | String | 100 chars | (empty) | Tertiary header text (promotional message, seasonal greeting, or blank) |
header_logo | Image URL | – | NULL | Uploaded logo image (max 300px wide; auto-scaled to paper width; monochrome recommended for thermal printers) |
Logo Specifications:
- Format: PNG or BMP (monochrome 1-bit BMP preferred for thermal printers).
- Maximum width: 300 pixels. Height auto-scales proportionally.
- The logo prints above
header_line_1. - If no logo is uploaded, the header begins with
header_line_1.
5.14.4 Footer Configuration
The receipt footer appears at the bottom of every printed receipt and supports up to 3 customizable text lines.
| Field | Type | Max Length | Default | Description |
|---|---|---|---|---|
footer_line_1 | String | 200 chars | “Returns accepted within 30 days with receipt.” | Primary footer text (typically return policy) |
footer_line_2 | String | 200 chars | (empty) | Secondary footer text (website URL, social media handles) |
footer_line_3 | String | 200 chars | “Thank you!” | Tertiary footer text (closing message) |
Business Rules:
- Footer lines can be blank (empty string). Blank lines are omitted from the printed receipt – no empty space is printed.
- Footer text should fit within the character-per-line limit of the configured paper width. Text exceeding the limit is word-wrapped automatically.
5.14.5 Receipt Configuration Data Model
Receipt Config Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
location_id | UUID | No | FK to locations table – NULL for tenant-wide default; non-NULL for location-specific override |
paper_width | Enum | Yes | 58MM, 80MM (default: 80MM) |
font_size | Enum | Yes | SMALL, MEDIUM, LARGE (default: MEDIUM) |
line_separator | Enum | Yes | DASH, EQUALS, BLANK, STAR (default: DASH) |
alignment | Enum | Yes | LEFT, CENTER (default: CENTER) |
print_density | Enum | Yes | LIGHT, NORMAL, DARK (default: NORMAL) |
header_lines | JSON | Yes | {"line_1": "...", "line_2": "...", "line_3": "..."} |
footer_lines | JSON | Yes | {"line_1": "...", "line_2": "...", "line_3": "..."} |
header_logo_url | String(500) | No | URL to uploaded header logo image |
field_order | JSON Array | Yes | Ordered array of field_code strings defining print sequence |
show_fields | JSON Object | Yes | Map of field_code: boolean controlling visibility (e.g., {"store_name": true, "register_number": false, ...}) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Configuration Inheritance:
- The system first looks for a location-specific receipt configuration (
location_id = target location). - If no location-specific configuration exists, it falls back to the tenant-wide default (
location_id = NULL). - Every tenant is initialized with a tenant-wide default receipt configuration using the default values from the field registry.
5.14.6 Email Receipt Template
Email receipts are HTML-formatted and sent when a customer provides an email address at checkout or explicitly requests an email receipt.
Email Receipt Template Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
subject_line | String(200) | Yes | Email subject (default: “Your receipt from {{store_name}}”) |
html_template | Text | Yes | HTML template body with merge fields |
is_active | Boolean | Yes | Whether email receipts are enabled (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Supported Merge Fields:
| Merge Field | Resolves To |
|---|---|
{{store_name}} | Location name |
{{store_address}} | Location full address |
{{store_phone}} | Location phone number |
{{transaction_id}} | Transaction number |
{{transaction_date}} | Formatted date and time |
{{items}} | HTML table of line items (name, qty, price, discount, total) |
{{subtotal}} | Pre-tax subtotal |
{{discount_total}} | Total discounts applied |
{{tax_amount}} | Tax amount with rate |
{{total}} | Grand total |
{{payment_method}} | Payment method(s) used |
{{customer_name}} | Customer name (if attached) |
{{loyalty_points_earned}} | Points earned on this transaction |
{{loyalty_balance}} | Current loyalty point balance |
{{barcode_image}} | Inline barcode image of transaction number |
Email Receipt Business Rules:
- Email receipts use the tenant’s branding colors (from Section 5.2) for header background, button colors, and accent elements.
- The company logo from the receipt header configuration is placed at the top of the email template.
- Every email receipt includes an unsubscribe link at the bottom: “Unsubscribe from receipt emails.” Clicking this sets the customer’s
email_receipt_opt_out = true. - Email receipts are queued asynchronously – the POS terminal does not wait for email delivery confirmation before completing the transaction.
- If the email fails to send (invalid address, mailbox full), the failure is logged but does not affect the transaction. The staff member sees no error; the customer simply does not receive the email.
- Email receipts are retained in the system for 7 years for audit and compliance purposes.
5.14.7 Receipt Preview
The Admin Portal provides a live preview of the receipt configuration, rendering a sample receipt with placeholder data so the administrator can verify layout, field order, and branding before saving.
Preview Behavior:
- Preview updates in real-time as the administrator toggles fields, reorders sections, or modifies header/footer text.
- Preview renders at the configured paper width (58mm or 80mm) using a monospace font to simulate thermal printer output.
- A “Send Test Email” button sends a sample email receipt to the administrator’s email address using the current email template configuration.
Cross-Reference: See Module 5, Section 5.8 for receipt printer hardware configuration and register-printer linking. See Module 1 for the transaction completion flow that triggers receipt printing.
5.15 Email Templates & Communications
Scope: Centralized email provider configuration and template registry for all automated communications sent by the POS system. This section covers SMTP/API provider setup, the complete email template catalog consolidated from all modules, merge field definitions, and per-template enablement controls.
Cross-Reference: See Module 1, Section 1.13 for sales-triggered email events. See Module 2, Section 2.9 for customer communication preferences. See Module 4, Section 4.16 for inventory alert email templates.
5.15.1 Email Provider Configuration
MOVED TO MODULE 6: Email provider configuration, data model, and business rules have been consolidated into Module 6, Section 6.9.1 (Provider Configuration).
See: Module 6, Section 6.9 for the complete email provider integration specification including SMTP/SendGrid/Mailgun configuration and delivery monitoring.
5.15.2 Template Registry
Every automated email sent by the POS system is defined as a template in a central registry. Templates are pre-seeded during tenant onboarding and can be individually enabled or disabled by the tenant administrator.
Consolidated Email Template Catalog
| Template Code | Source Module | Trigger Event | Default Recipients | Description |
|---|---|---|---|---|
TMPL-REFUND-CONFIRMATION | Sales (M1) | Refund processed | Customer email | Confirms refund amount, method, and expected processing time |
TMPL-SPECIAL-ORDER-READY | Sales (M1) | Special order arrived | Customer email | Notifies customer that their special order is ready for pickup |
TMPL-SHIPMENT-TRACKING | Sales (M1) | Ship-to-customer dispatched | Customer email | Provides carrier name and tracking number |
TMPL-DELIVERY-CONFIRMATION | Sales (M1) | Delivery confirmed | Customer email | Confirms package delivery with order summary |
TMPL-OFFLINE-SOLD | Sales (M1) | Reserved item sold offline | Customer email | Informs customer when an offline-sold transfer/ship/reserve item is unavailable |
TMPL-WELCOME | Customers (M2) | New customer created | Customer email | Welcome message with loyalty program introduction |
TMPL-TIER-UPGRADE | Customers (M2) | Loyalty tier change | Customer email | Congratulates customer on tier upgrade with new benefits summary |
TMPL-PO-VENDOR | Inventory (M4) | PO submitted to vendor | Vendor email | Formatted purchase order with line items, quantities, and expected delivery |
TMPL-TRANSFER-ALERT | Inventory (M4) | Transfer shipped | Destination manager | Notifies destination store that a transfer is in transit |
TMPL-LOW-STOCK | Inventory (M4) | Daily low stock digest | Store manager | Consolidated list of products below reorder point at each location |
TMPL-COUNT-REMINDER | Inventory (M4) | Upcoming count scheduled | Assigned counters | Reminder email with count date, location, and scope (full/cycle) |
TMPL-RECEIPT-EMAIL | Setup (M5) | Customer requests email receipt | Customer email | Full transaction receipt in HTML format with scannable barcode |
TMPL-PASSWORD-RESET | Setup (M5) | User password reset request | User email | Secure password reset link with 24-hour expiry |
5.15.3 Template Data Model
Email Template Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
code | String(50) | Yes | Unique template code (e.g., TMPL-REFUND-CONFIRMATION). Immutable after creation. |
name | String(100) | Yes | Human-readable template name (e.g., “Refund Confirmation”) |
subject_template | String(255) | Yes | Email subject with merge fields (e.g., “Your refund of {refund_amount} has been processed”) |
body_template | Text | Yes | HTML email body with merge fields. Supports inline CSS for styling. Maximum 50KB. |
trigger_event | String(100) | Yes | System event that triggers this email (e.g., REFUND_PROCESSED, SPECIAL_ORDER_ARRIVED) |
default_recipient_type | Enum | Yes | CUSTOMER, VENDOR, STORE_MANAGER, ASSIGNED_USER, CUSTOM |
custom_recipient_email | String(255) | No | Static email address when default_recipient_type = CUSTOM |
is_enabled | Boolean | Yes | Whether this template is active and will be sent when triggered (default: true) |
is_system | Boolean | Yes | true for pre-seeded templates, false for tenant-created (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
5.15.4 Merge Fields
Each template uses merge fields enclosed in curly braces. The system resolves merge fields at send time by injecting context-specific values from the triggering event.
Common Merge Fields (available in all templates):
| Merge Field | Type | Description |
|---|---|---|
{tenant_name} | String | Tenant trading name |
{store_name} | String | Location name where the event occurred |
{store_address} | String | Full store address |
{store_phone} | String | Store phone number |
{current_date} | Date | Date the email is generated |
{current_time} | Time | Time the email is generated |
Transaction Merge Fields (sales-triggered templates):
| Merge Field | Type | Description |
|---|---|---|
{customer_name} | String | Customer full name |
{transaction_id} | String | Transaction reference number |
{order_total} | Currency | Total transaction amount |
{refund_amount} | Currency | Refund amount (refund templates only) |
{payment_method} | String | Payment method used |
{tracking_number} | String | Carrier tracking number (shipment templates only) |
{carrier_name} | String | Shipping carrier name (shipment templates only) |
{pickup_deadline} | Date | Pickup deadline date (special order and hold templates) |
Inventory Merge Fields (inventory-triggered templates):
| Merge Field | Type | Description |
|---|---|---|
{po_number} | String | Purchase order number |
{vendor_name} | String | Vendor company name |
{transfer_number} | String | Transfer reference number |
{source_location} | String | Transfer origin location name |
{destination_location} | String | Transfer destination location name |
{count_date} | Date | Scheduled count date |
{count_type} | String | Count type (Full Physical, Cycle, On-Demand, etc.) |
{low_stock_items} | HTML Table | Rendered table of low-stock products (digest templates only) |
Business Rules:
- Unresolved merge fields are replaced with an empty string and logged as a warning. They do not prevent the email from sending.
- Tenant administrators can customize the
subject_templateandbody_templateof system templates but cannot modify thecode,trigger_event, ordefault_recipient_type. - Disabling a template (
is_enabled = false) prevents the email from being sent when the trigger event fires. The event itself still processes normally. - Email receipts (
TMPL-RECEIPT-EMAIL) render the same field data as printed receipts, formatted in responsive HTML with a scannable CODE-128 barcode image.
5.16 RFID Configuration
Scope: Configuration of RFID hardware, EPC encoding parameters, tag printing, and scan session settings for the dedicated inventory counting subsystem. RFID is a counting-only system — it counts inventory through bulk tag reads. It does NOT participate in sales transactions, receiving, or transfers. Barcode Scanners remain the input device for those workflows (see Section 4.4 Receiving, Section 1.A.1 Item Entry).
Terminology Distinction:
- Scanner = barcode input device used at the POS register for sales, returns, receiving, and item lookup (Modules 1, 3, 4). Operates one-item-at-a-time via USB HID keyboard wedge.
- RFID = dedicated counting subsystem using radio-frequency readers for bulk inventory counting and auditing (Module 4, Section 4.6). Operates via the Raptag mobile application, reading 40+ tags per second.
- These are separate abstractions that coexist. Decision #11 (“Scanner Terminology”) applies to barcode input. RFID has its own configuration and workflow documented here.
Cross-References:
- Chapter 16 (Raptag Mobile Application) — mobile RFID counting interface
- Section 4.6.8 (RFID-Assisted Counting) — counting workflow integration
- Module 6, Section 6.11 (Integration Hub) — external system integrations (Shopify, Amazon, Google)
5.16.1 Reader Registration
RFID readers are registered as enterprise devices, paired to a location, and managed via claim codes generated in the Admin Portal.
Supported Reader Models:
| Model | Form Factor | Read Range | Use Case | Connectivity |
|---|---|---|---|---|
| Zebra MC3390R | Handheld gun | 20 ft | Full store inventory counts | WiFi, Bluetooth |
| Zebra RFD40 | Phone sled attachment | 12 ft | Zone/section counts | Bluetooth |
| Zebra FX9600 | Fixed (dock door) | 30 ft | Receiving dock verification | Ethernet, WiFi |
Reader Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to shared.tenants |
name | VARCHAR(100) | Yes | Human-readable name (e.g., “Store GM Handheld #1”) |
model | VARCHAR(50) | Yes | Reader model (MC3390R, RFD40, FX9600) |
serial_number | VARCHAR(50) | Yes | Manufacturer serial number |
location_id | INT | Yes | FK to locations — reader’s assigned location |
connection_type | VARCHAR(20) | Yes | wifi, bluetooth, ethernet |
claim_code | VARCHAR(6) | No | One-time registration code (e.g., “X7K9M2”) |
status | VARCHAR(20) | Yes | active, offline, maintenance, retired |
registered_at | TIMESTAMP | Yes | When device was first registered |
last_seen_at | TIMESTAMP | No | Last heartbeat from device |
Claim Code Registration Workflow:
- Admin generates a 6-character alphanumeric claim code in Admin Portal (
Settings > RFID > Devices) - Claim code is valid for 24 hours from generation
- Operator enters claim code on the Raptag mobile app during setup
- System validates claim code, registers device to the tenant and location
- Claim code is consumed (one-time use)
Business Rules:
- One reader can be registered to one location at a time
- A reader can be reassigned to a different location by an Admin
- Readers with
status = 'offline'for >15 minutes trigger an alert in the Admin Portal - Retired readers cannot be re-activated — a new claim code must be generated
5.16.2 EPC Encoding Configuration
All RFID tags use the SGTIN-96 standard (96-bit EPC, encoded as 24 hexadecimal characters). Each tenant configures their EPC encoding parameters once during onboarding.
SGTIN-96 Structure:
Header (8 bits) | Filter (3) | Partition (3) | Company Prefix (20-40) | Item Ref (4-24) | Serial (38)
Tenant EPC Configuration:
| Field | Type | Default | Description |
|---|---|---|---|
epc_company_prefix | VARCHAR(24) | — | GS1-assigned company prefix (set during onboarding) |
epc_indicator | CHAR(1) | 0 | SGTIN indicator digit |
epc_filter | CHAR(1) | 3 | Filter value (3 = individual trade item, per GS1 spec) |
epc_partition | INT | 5 | Partition value (5 = 20-bit company prefix + 24-bit item reference) |
min_rssi_threshold | SMALLINT | -70 | Minimum RSSI in dBm to accept a tag read; weaker reads are filtered as phantom reads |
Serial Number Management:
- Serial numbers use a PostgreSQL SEQUENCE per tenant (not a column counter)
- Sequence:
CREATE SEQUENCE rfid_epc_serial_{tenant_short_id} START 1 INCREMENT 1 NO CYCLE - Application calls
nextval()during tag encoding — guarantees uniqueness under concurrent printing - 38-bit serial field supports up to 274 billion unique tags per company prefix
EPC Format Validation:
- All EPCs must match:
^[0-9A-F]{24}$(exactly 24 uppercase hexadecimal characters) - Enforced via database CHECK constraint on
rfid_tags.epc
Scope Constraint: RFID is counting-only. The rfid_tags table tracks tag status as active, void, or lost. There are no sold_at, transferred_at, or sold_order_id fields — sales and transfers are tracked by the core inventory system via barcode, not RFID.
5.16.3 Tag Printing Parameters
RFID tags are printed and encoded using dedicated RFID-enabled label printers. The POS system manages the full tag printing lifecycle: template design, job queue, encoding, and verification.
Supported Printer Models:
| Model | Manufacturer | DPI | Connection | RFID Position |
|---|---|---|---|---|
| ZD621R | Zebra | 300 | Network, USB | Center |
| ZD500R | Zebra | 203 | Network, USB, Bluetooth | Center |
| CL4NX | SATO | 305 | Network, USB | Left |
| MX240P | TSC | 203 | Network, USB | Right |
Template Types:
| Type | Use Case | Typical Size |
|---|---|---|
hang_tag | Clothing hang tags with price and size | 2“ x 3“ |
price_tag | Shelf price labels with EPC | 1.5“ x 1“ |
label | Adhesive labels for boxes/bins | 4“ x 2“ |
Templates use ZPL (Zebra Programming Language) format. SATO and TSC printers accept ZPL via built-in translation.
Print Job Configuration:
| Setting | Default | Description |
|---|---|---|
default_priority | 5 | Job priority (1=highest, 10=lowest) |
max_retry_attempts | 3 | Retries for failed tag encoding |
job_timeout_minutes | 30 | Max time before job is marked failed |
| Default printer per location | — | Set in Admin Portal > RFID > Printers |
Business Rules:
- Large print jobs (>1,000 tags) should be split into sub-jobs of 500-1,000 tags for progress tracking
- Failed tags within a job can be retried individually without resubmitting the entire job
- If the assigned printer goes offline mid-job, the job pauses until the printer recovers (no automatic failover in v1.0)
5.16.4 Scan Session Configuration
RFID scan sessions are the core counting operation. A session represents a single counting activity — from starting the reader to submitting results.
Session Types (Counting Only):
| Type Code | Name | Scope | Typical Items |
|---|---|---|---|
full_inventory | Full Store Count | All products at a location | 2,000 – 100,000+ |
cycle_count | Cycle Count | Rolling partial count by category | 200 – 2,000 |
spot_check | Spot Check | Discrepancy verification | 10 – 50 |
find_item | Find Item | Locate a specific SKU using reader | 1 |
Note:
receivingis NOT an RFID session type. Receiving uses barcode scanners (Section 4.4).
Session Parameters:
| Setting | Key | Default | Description |
|---|---|---|---|
| Session Timeout | session_timeout_minutes | 480 (8 hours) | Max session duration before auto-expire |
| Auto-Save Interval | auto_save_interval_seconds | 30 | Frequency of SQLite checkpoint writes on mobile device |
| Chunk Upload Size | chunk_upload_size | 5,000 | Events per upload chunk when syncing to server |
| RSSI Threshold | min_rssi_threshold | -70 dBm | Tag reads below this are filtered (phantom read prevention) |
Variance Thresholds:
| Variance % | Color | Action |
|---|---|---|
| 0% | Green | Auto-approve — no discrepancy |
| 1–2% | Yellow | Review recommended |
| 3–5% | Orange | Manager review required |
| >5% | Red | Mandatory recount with different operator |
Multi-Operator Support:
- A single session can have multiple operators (up to 10), each assigned to a section of the store
- Each operator scans independently using their own device
- Server merges results and deduplicates by EPC (keeps highest RSSI read)
- See Section 4.6.8 for the full multi-operator workflow
5.16.5 RFID Business Rules (YAML)
The following RFID-specific business rules are part of the consolidated configuration system (Section 5.19). They are documented here for reference and cross-referenced from the YAML block.
rfid_config:
# EPC Encoding
epc:
company_prefix: "" # Tenant-specific, set during onboarding
partition: 5 # 20-bit company prefix + 24-bit item reference
filter: 3 # Individual trade item (GS1 standard)
indicator: "0" # SGTIN indicator digit
format: "SGTIN-96" # 96-bit EPC standard
serial_strategy: "sequence" # PostgreSQL SEQUENCE (not column counter)
format_regex: "^[0-9A-F]{24}$"
# Reader Hardware
readers:
supported_models:
- "MC3390R" # Handheld gun, 20 ft range
- "RFD40" # Phone sled, 12 ft range
- "FX9600" # Fixed reader, 30 ft range
claim_code_length: 6
claim_code_expiry_hours: 24
heartbeat_interval_seconds: 300
offline_threshold_minutes: 15
# Scanning Sessions
scanning:
session_types:
- "full_inventory"
- "cycle_count"
- "spot_check"
- "find_item"
# NOTE: "receiving" is NOT an RFID session type
min_rssi_threshold: -70 # dBm, tags weaker than this are filtered
auto_save_interval_seconds: 30
session_timeout_minutes: 480 # 8 hours max
chunk_upload_size: 5000 # Events per upload chunk
max_operators_per_session: 10
# Tag Printing
printing:
default_priority: 5 # 1=highest, 10=lowest
max_retry_attempts: 3
job_timeout_minutes: 30
max_tags_per_sub_job: 1000 # Large jobs split into sub-jobs
# Variance Thresholds
variance:
auto_approve_threshold_percent: 0
review_threshold_percent: 2
manager_review_threshold_percent: 5
recount_required_threshold_percent: 20
5.16.6 Integration Hub Reference
Note: The Integration Hub (integration registry, credentials storage, Shopify/Amazon/Google configurations, sync logging, and health dashboard) has been consolidated into Module 6, Section 6.11. RFID is a first-party subsystem and does NOT use the Integration Hub — it connects directly to the central API via REST endpoints documented in Chapter 18 and Appendix A.
5.17 Loyalty & Rewards Settings
Scope: Configurable parameters for the loyalty program, tier thresholds, reward redemption rates, and gift card settings. This section defines the settings – the configurable values that govern loyalty behavior. The loyalty rules (tier upgrade/downgrade logic, point accrual timing, redemption application in the payment flow) are defined in Module 2 (Customers).
Cross-Reference: See Module 2, Section 2.6 for loyalty tier upgrade/downgrade rules, point accrual logic, and redemption flow. See Module 1, Section 1.15 for loyalty redemption in the payment calculation sequence.
5.17.1 Point Configuration
| Setting | Key | Type | Default | Description |
|---|---|---|---|---|
| Base Earn Rate | points_per_dollar | Integer | 1 | Points earned per dollar spent (before tier multiplier). Applied to the post-tax total. |
| Points Expiry | points_expiry_months | Integer | 12 | Months after last earning activity before points expire. 0 = never expire. |
| Exclude Tax | exclude_tax_from_points | Boolean | false | If true, points are calculated on the pre-tax subtotal. |
| Exclude Discounted Amount | exclude_discounts_from_points | Boolean | false | If true, points are calculated on the original price, not the discounted price. |
5.17.2 Tier Thresholds
Tier definitions are configurable. The system supports up to 4 tiers. Each tier defines a spend threshold, point multiplier, and automatic discount percentage.
| Tier | Code | Annual Spend Threshold | Point Multiplier | Auto Discount % | Description |
|---|---|---|---|---|---|
| Bronze | BRONZE | $0 (default tier) | 1.0x | 0% | Entry tier – all new customers start here |
| Silver | SILVER | $1,000 | 1.5x | 5% | Mid-tier – 50% more points per dollar, 5% discount on all purchases |
| Gold | GOLD | $5,000 | 2.0x | 10% | Premium tier – double points, 10% automatic discount |
| Platinum | PLATINUM | $10,000 | 3.0x | 15% | Top tier – triple points, 15% automatic discount |
Business Rules:
- Tier thresholds are evaluated against the customer’s rolling 12-month spend total. The evaluation period resets annually on the customer’s enrollment anniversary date.
- The automatic discount is applied before any manual or promotional discounts in the calculation order (see Module 1 discount application order).
- Point multiplier applies to the base
points_per_dollarrate. A Gold customer earning 1 point per dollar receives 2 points per dollar.
Cross-Reference: See Module 2, Section 2.6 for tier upgrade trigger logic, downgrade grace period, and tier evaluation cadence.
5.17.3 Reward Redemption
| Setting | Key | Type | Default | Description |
|---|---|---|---|---|
| Redemption Rate | redemption_rate | Integer | 100 | Points required for $1.00 discount |
| Minimum Redemption | minimum_redemption | Integer | 100 | Minimum points that can be redeemed in a single transaction |
| Maximum Redemption % | max_redemption_percent | Integer | 50 | Maximum percentage of the transaction total payable by points (0-100) |
| Allow Partial Redemption | allow_partial_redemption | Boolean | true | Whether customers can redeem a subset of their available points |
5.17.4 Gift Card Settings
| Setting | Key | Type | Default | Description |
|---|---|---|---|---|
| Predefined Denominations | denominations | Array[Decimal] | [10, 25, 50, 100] | Quick-select amounts shown at POS during gift card purchase |
| Allow Custom Amount | allow_custom_amount | Boolean | true | Whether cashiers can enter an arbitrary gift card amount |
| Minimum Load | minimum_load | Decimal(10,2) | 10.00 | Minimum dollar amount for initial activation or reload |
| Maximum Load | maximum_load | Decimal(10,2) | 500.00 | Maximum dollar amount for initial activation or reload |
| Expiry Months | expiry_months | Integer | 0 | Months from activation before the gift card expires. 0 = no expiry (most restrictive jurisdiction default). |
| Allow Reload | allow_reload | Boolean | true | Whether depleted or partially-used gift cards can be reloaded |
Jurisdiction Rules:
- Default expiry is
0(no expiry), conforming to the most restrictive US jurisdiction (California). - Tenants operating in states that permit expiry can override
expiry_monthsto a compliant value (e.g., Virginia: minimum 60 months). - Jurisdiction-specific cash-out rules (e.g., California requires cash redemption below $10.00) are enforced at the POS transaction level per Module 1 gift card rules.
5.17.5 Loyalty Settings Data Model
Loyalty Settings Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant (unique constraint) |
points_per_dollar | Integer | Yes | Base earn rate (default: 1) |
points_expiry_months | Integer | Yes | Months to expiry, 0 = never (default: 12) |
exclude_tax_from_points | Boolean | Yes | Default: false |
exclude_discounts_from_points | Boolean | Yes | Default: false |
redemption_rate | Integer | Yes | Points per $1.00 discount (default: 100) |
minimum_redemption | Integer | Yes | Minimum redeemable points (default: 100) |
max_redemption_percent | Integer | Yes | Max % of transaction payable by points (default: 50) |
allow_partial_redemption | Boolean | Yes | Default: true |
tier_config | JSON | Yes | JSON array of tier objects: [{code, name, threshold, multiplier, discount_percent}] |
gift_card_denominations | JSON | Yes | JSON array of decimal amounts (default: [10, 25, 50, 100]) |
gift_card_allow_custom | Boolean | Yes | Default: true |
gift_card_min_load | Decimal(10,2) | Yes | Default: 10.00 |
gift_card_max_load | Decimal(10,2) | Yes | Default: 500.00 |
gift_card_expiry_months | Integer | Yes | Default: 0 |
gift_card_allow_reload | Boolean | Yes | Default: true |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
5.18 Audit Configuration
Scope: Configuration of the audit logging system – which event categories are tracked, how long logs are retained, archival policies, and export capabilities. The audit log provides a tamper-evident record of every significant action in the system for compliance, investigation, and operational accountability.
Cross-Reference: See Module 5, Section 5.13 for approval workflow audit trails. See Module 4, Section 4.9 for inventory movement history (the inventory-specific audit trail).
5.18.1 Audit Categories
Each audit category can be independently toggled on or off. Disabling a category stops new log entries from being created for events in that category. Existing log entries are never deleted by disabling a category.
| Category Code | Description | Default | Example Events |
|---|---|---|---|
LOGIN | User login and logout events | On | Login success, login failure, logout, session timeout |
SALE | Transaction completed | On | Sale finalized, split payment processed |
RETURN | Return processed | On | Return with receipt, return without receipt, exchange |
VOID | Transaction voided | On | Same-day void, void with manager approval |
ADJUSTMENT | Inventory adjustment | On | Manual qty adjustment, count correction applied |
PRICE_CHANGE | Product price modified | On | Retail price change, cost update, markdown applied |
DISCOUNT | Discount applied | On | Line discount, global discount, coupon applied, loyalty redemption |
PO | Purchase order actions | On | PO created, PO approved, PO submitted, PO received, PO closed |
TRANSFER | Inter-store transfer actions | On | Transfer requested, approved, shipped, received, completed |
USER_MGMT | User account actions | On | User created, role changed, user deactivated, password reset |
SETTINGS | System settings changed | On | Tax rate changed, business rule modified, integration updated |
INVENTORY_COUNT | Count session actions | On | Count started, count submitted, variance approved, count finalized |
5.18.2 Retention Configuration
| Setting | Key | Type | Default | Description |
|---|---|---|---|---|
| Retention Period | retention_days | Integer | 365 | Days to keep detailed audit log entries in the primary database |
| Archive Enabled | archive_enabled | Boolean | true | Whether entries older than retention_days are moved to archive storage |
| Archive Format | archive_format | Enum | COMPRESSED_JSON | Format for archived records: COMPRESSED_JSON (gzip), CSV |
| Purge Archived | purge_archived_after_days | Integer | 2190 | Days to retain archived records before permanent deletion (2190 = 6 years). Set to 0 to retain archives indefinitely. |
Retention Lifecycle:
flowchart LR
A["Active Log\n(Primary DB)"] -->|After retention_days| B["Archive Storage\n(Compressed JSON)"]
B -->|After purge_archived_after_days| C["Permanently Deleted"]
style A fill:#2d6a4f,stroke:#1b4332,color:#fff
style B fill:#264653,stroke:#1d3557,color:#fff
style C fill:#6c757d,stroke:#495057,color:#fff
Business Rules:
- Minimum
retention_daysis 90. The system rejects values below 90 to ensure basic operational audit capability. - Minimum
purge_archived_after_daysis 365 (or 0 for indefinite). Values between 1 and 364 are rejected. - The archival background job runs daily at 02:00 AM (tenant timezone). It processes entries older than
retention_daysin batches of 10,000 records. - Archived records are stored with the same data fidelity as active records – no fields are dropped during archival.
5.18.3 Export Configuration
| Setting | Key | Type | Default | Description |
|---|---|---|---|---|
| Supported Formats | export_formats | Array[Enum] | ["CSV", "JSON", "PDF"] | Formats available for audit log export |
| Max Export Rows | max_export_rows | Integer | 10000 | Maximum rows per single export request. Larger exports must be split by date range. |
| Max Date Range | max_export_date_range_days | Integer | 365 | Maximum date range span allowed in a single export request |
| Include Archived | include_archived_in_export | Boolean | true | Whether exports can pull from archived records (slower but comprehensive) |
5.18.4 Audit Log Data Model
Audit Config Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant (unique constraint) |
categories_enabled | JSON | Yes | Map of category_code: boolean (e.g., {"LOGIN": true, "SALE": true, "VOID": true, ...}) |
retention_days | Integer | Yes | Default: 365 |
archive_enabled | Boolean | Yes | Default: true |
archive_format | Enum | Yes | COMPRESSED_JSON, CSV (default: COMPRESSED_JSON) |
purge_archived_after_days | Integer | Yes | Default: 2190 |
max_export_rows | Integer | Yes | Default: 10000 |
max_export_date_range_days | Integer | Yes | Default: 365 |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Audit Log Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
category_code | Enum | Yes | One of the 12 category codes defined above |
action | String(100) | Yes | Specific action (e.g., SALE_COMPLETED, USER_CREATED, PO_APPROVED) |
actor_user_id | UUID | Yes | FK to users table – user who performed the action |
actor_role | String(50) | Yes | Role of the user at time of action (captured for audit, not FK) |
location_id | UUID | No | FK to locations table – location where the action occurred (NULL for tenant-wide actions) |
register_id | UUID | No | FK to registers table – register involved (NULL for non-POS actions) |
entity_type | String(50) | Yes | Type of entity affected (e.g., TRANSACTION, PRODUCT, USER, PO) |
entity_id | UUID | Yes | FK to the affected entity record |
details | JSON | No | Structured JSON capturing before/after values, amounts, reason codes, and other context |
ip_address | String(45) | No | IP address of the client that initiated the action |
occurred_at | DateTime | Yes | Timestamp when the action occurred (event time, not log write time) |
created_at | DateTime | Yes | Record creation timestamp (log write time) |
Business Rules:
- Audit log entries are immutable. No UPDATE or DELETE operations are permitted on the
audit_logtable. Only INSERT and SELECT are allowed. - The
detailsJSON field captures change context. For settings changes, it records{"old_value": ..., "new_value": ...}. For transactions, it records key financial fields (total,tax,discount,payment_method). - Audit logs are queryable by category, date range, user, location, and entity type. The Admin Portal provides a searchable, filterable audit log viewer.
5.19 Business Rules Configuration – Consolidated YAML
Scope: Single, authoritative source for all configurable business rules across the entire POS system. This section consolidates business rules from Module 1 (Sales), Module 2 (Customers), Module 3 (Catalog), and Module 4 (Inventory) into one centralized YAML configuration. All values shown are defaults and can be overridden at tenant or store level.
Cross-Reference: Individual module sections define the behavior that these rules govern. This section defines the configurable values.
5.19.1 Sales Configuration
# ============================================
# SALES MODULE BUSINESS RULES
# ============================================
# All values shown are defaults and can be
# overridden at tenant or store level.
# ============================================
sales_config:
# ------------------------------------------
# RETURN POLICY
# ------------------------------------------
return_policy:
# Full refund period (with receipt)
full_refund_days: 30
# Store credit only period (with receipt)
store_credit_days: 90
# Restocking fee for opened items (percentage)
restocking_fee_percent: 15
# Categories exempt from restocking fee
restocking_fee_exempt_categories:
- "clothing"
- "accessories"
# Categories marked as final sale (no returns)
final_sale_categories:
- "clearance"
- "as-is"
# Channel-specific policies
online_policy:
return_days: 30
exchange_days: 30
exclude_shipping_fees: true
exclude_processing_fees: true
receipt_required: true
in_store_policy:
return_hours: 24
exchange_hours: 24
receipt_required: true
receipt_scan_validation: true
# ------------------------------------------
# PARKED SALES
# ------------------------------------------
parked_sales:
# Maximum parked sales per terminal
max_per_terminal: 5
# Time-to-live before auto-expiry (hours)
ttl_hours: 4
# Inventory reservation type
reservation_type: "soft" # soft = visible with warning, hard = blocked
# ------------------------------------------
# SPECIAL ORDERS
# ------------------------------------------
special_orders:
# Minimum deposit percentage
minimum_deposit_percent: 50
# Maximum days to hold after arrival
pickup_deadline_days: 30
# Auto-cancel after missed pickup (days)
abandonment_days: 45
# ------------------------------------------
# HOLD FOR PICKUP
# ------------------------------------------
hold_for_pickup:
# Default hold duration
default_days: 7
# Maximum hold extension allowed
max_days: 30
# Days before expiry to send reminder
reminder_days_before: 2
# ------------------------------------------
# TRANSFERS & RESERVATIONS
# ------------------------------------------
transfers:
# Require full payment before processing
require_full_payment: true
# Estimated transit days (for display)
estimated_transit_days: 3
reservations:
# Require full payment before reserving
require_full_payment: true
# Default reservation duration
default_hold_days: 7
# Auto-refund after expiry
auto_refund_on_expiry: true
# ------------------------------------------
# SHIP TO CUSTOMER
# ------------------------------------------
ship_to_customer:
enabled: true
require_full_payment: true
include_shipping_in_total: true
# Carrier integration
carriers:
- provider: "configured_per_tenant"
api_key: "configured_per_tenant"
test_mode: true
# Shipping options
shipping_options:
standard:
label: "Standard (3-5 business days)"
enabled: true
express:
label: "Express (1-2 business days)"
enabled: true
overnight:
label: "Overnight"
enabled: false
# ------------------------------------------
# CASH DRAWER
# ------------------------------------------
cash_drawer:
# Variance tolerance before manager approval required
variance_tolerance: 5.00
# Require blind count (staff can't see expected)
blind_count_enabled: true
# Maximum float amount
max_opening_float: 500.00
# ------------------------------------------
# DISCOUNTS & PRICING
# ------------------------------------------
discounts:
# Maximum line item discount without manager approval
max_line_discount_percent: 20
# Maximum global discount without manager approval
max_global_discount_percent: 15
# Reason codes required for discounts
require_reason_code: true
# Discount application order
application_order:
- "price_tier"
- "line_discount"
- "auto_promo"
- "global_discount"
- "coupon"
- "tax"
- "loyalty_redemption"
# ------------------------------------------
# COMMISSIONS
# ------------------------------------------
commissions:
# Default commission rate (percentage of sale)
default_rate_percent: 2.0
# Higher rate for specific categories
category_rates:
electronics: 3.0
services: 5.0
# Void reverses commission (full)
reverse_on_void: true
void_reversal_method: "full"
# Return reduces commission (proportional)
reduce_on_return: true
return_reversal_method: "proportional"
# ------------------------------------------
# OFFLINE MODE
# ------------------------------------------
offline_mode:
# Maximum transactions to queue locally
max_queue_size: 100
# Auto-sync interval when online (seconds)
sync_interval_seconds: 30
# Conflict resolution strategy
conflict_strategy: "server_wins_with_review"
# Operations allowed offline
allowed_offline:
- "sale_new"
- "return_with_receipt"
- "price_check"
- "parked_sale_create"
- "parked_sale_retrieve"
# Operations blocked offline
blocked_offline:
- "customer_create"
- "on_account_payment"
- "gift_card_activation"
- "gift_card_reload"
- "multi_store_inventory"
- "transfer_request"
- "reservation_create"
# ------------------------------------------
# PAYMENT INTEGRATION
# ------------------------------------------
payment_integration:
# PCI compliance level
pci_scope: "SAQ-A"
# Integration type
integration_type: "semi_integrated"
# Terminal timeout (seconds)
payment_timeout_seconds: 60
connection_timeout_seconds: 10
# Batch close time (24-hour format)
batch_close_time: "23:00"
# Same-day void allowed
same_day_void: true
# ------------------------------------------
# THIRD-PARTY FINANCING
# ------------------------------------------
third_party_financing:
affirm:
enabled: true
minimum_order_amount: 50.00
maximum_order_amount: 5000.00
merchant_id: "configured_per_tenant"
webhook_url: "/api/webhooks/affirm"
Sales Configuration Field Reference
| Key | Type | Default | Description |
|---|---|---|---|
return_policy.full_refund_days | Integer | 30 | Calendar days from purchase for full original-method refund |
return_policy.store_credit_days | Integer | 90 | Calendar days from purchase for store credit refund |
return_policy.restocking_fee_percent | Integer | 15 | Percentage deducted as restocking fee on non-exempt items |
parked_sales.max_per_terminal | Integer | 5 | Maximum concurrent parked sales per register |
parked_sales.ttl_hours | Integer | 4 | Hours before an unrecalled parked sale auto-expires |
special_orders.minimum_deposit_percent | Integer | 50 | Minimum deposit required at special order creation |
special_orders.abandonment_days | Integer | 45 | Days after arrival before abandoned special order is forfeited |
hold_for_pickup.default_days | Integer | 7 | Default hold duration in days |
cash_drawer.variance_tolerance | Decimal | 5.00 | Dollar variance before manager review is required at close |
cash_drawer.blind_count_enabled | Boolean | true | Hide expected drawer total from staff during count |
discounts.max_line_discount_percent | Integer | 20 | Maximum per-line discount before manager approval |
discounts.max_global_discount_percent | Integer | 15 | Maximum transaction-wide discount before manager approval |
commissions.default_rate_percent | Decimal | 2.0 | Default commission percentage of sale amount |
offline_mode.max_queue_size | Integer | 100 | Maximum offline-queued transactions before blocking new sales |
payment_integration.payment_timeout_seconds | Integer | 60 | Seconds to wait for payment terminal response |
payment_integration.batch_close_time | String | “23:00” | Daily batch settlement time (24-hour format) |
5.19.2 Customer Configuration
# ============================================
# CUSTOMER MODULE BUSINESS RULES
# ============================================
customer_config:
# ------------------------------------------
# LOYALTY PROGRAM
# ------------------------------------------
loyalty:
# Points per dollar spent
points_per_dollar: 1
# Points required for $1 redemption
redemption_rate: 100
# Minimum points to redeem
minimum_redemption: 100
# Points expiry (months, 0 = never)
points_expiry_months: 12
# Maximum % of transaction payable by points
max_redemption_percent: 50
# ------------------------------------------
# CUSTOMER TIERS
# ------------------------------------------
customer_tiers:
bronze:
# No threshold - default tier
point_multiplier: 1.0
automatic_discount_percent: 0
silver:
annual_spend_threshold: 1000
point_multiplier: 1.5
automatic_discount_percent: 5
gold:
annual_spend_threshold: 5000
point_multiplier: 2.0
automatic_discount_percent: 10
platinum:
annual_spend_threshold: 10000
point_multiplier: 3.0
automatic_discount_percent: 15
# ------------------------------------------
# LAYAWAY
# ------------------------------------------
layaway:
# Minimum deposit percentage
minimum_deposit_percent: 20
# Maximum layaway duration (days)
max_duration_days: 90
# Minimum payment frequency (days)
payment_frequency_days: 30
# Forfeiture fee percentage (of deposit)
forfeiture_fee_percent: 10
# ------------------------------------------
# GIFT CARDS
# ------------------------------------------
gift_cards:
# Minimum load amount
minimum_load: 10.00
# Maximum load amount
maximum_load: 500.00
# Default expiry period (months from activation)
default_expiry_months: 0 # No expiry (most restrictive)
# Allow reload of depleted cards
allow_reload: true
# Jurisdiction-specific overrides
jurisdiction_rules:
virginia:
expiry_allowed: true
expiry_months: 60
inactivity_fee_allowed: true
inactivity_fee_months: 12
cash_out_threshold: 0
california:
expiry_allowed: false
inactivity_fee_allowed: false
cash_out_threshold: 10.00
cash_out_required: true
# ------------------------------------------
# PRIVACY & DATA RETENTION
# ------------------------------------------
privacy:
# Days to fulfill data export request
export_request_days: 30
# Days to fulfill deletion request
deletion_request_days: 30
# Auto-anonymize inactive customers (months, 0 = never)
auto_anonymize_months: 0
# Consent change audit retention (years)
consent_audit_retention_years: 7
# ------------------------------------------
# EXPORT LIMITS
# ------------------------------------------
exports:
# Maximum rows per CSV export
max_rows: 1000
# Maximum date range for reports (days)
max_date_range_days: 365
Customer Configuration Field Reference
| Key | Type | Default | Description |
|---|---|---|---|
loyalty.points_per_dollar | Integer | 1 | Base points earned per dollar (before tier multiplier) |
loyalty.redemption_rate | Integer | 100 | Points required to redeem $1.00 discount |
loyalty.minimum_redemption | Integer | 100 | Minimum points a customer can redeem per transaction |
loyalty.points_expiry_months | Integer | 12 | Months since last earn before points expire (0 = never) |
loyalty.max_redemption_percent | Integer | 50 | Maximum percentage of transaction total payable by points |
customer_tiers.silver.annual_spend_threshold | Integer | 1000 | Annual spend in dollars to qualify for Silver tier |
customer_tiers.gold.annual_spend_threshold | Integer | 5000 | Annual spend in dollars to qualify for Gold tier |
customer_tiers.platinum.annual_spend_threshold | Integer | 10000 | Annual spend in dollars to qualify for Platinum tier |
layaway.minimum_deposit_percent | Integer | 20 | Minimum deposit as percentage of layaway total |
layaway.max_duration_days | Integer | 90 | Maximum calendar days for layaway completion |
layaway.forfeiture_fee_percent | Integer | 10 | Fee deducted from deposit on layaway cancellation |
gift_cards.minimum_load | Decimal | 10.00 | Minimum dollar amount per activation or reload |
gift_cards.maximum_load | Decimal | 500.00 | Maximum dollar amount per activation or reload |
gift_cards.default_expiry_months | Integer | 0 | Months to expiry (0 = never, California baseline) |
privacy.export_request_days | Integer | 30 | Days to fulfill customer data export request |
privacy.auto_anonymize_months | Integer | 0 | Months of inactivity before auto-anonymization (0 = disabled) |
5.19.3 Catalog Configuration
# ============================================
# CATALOG MODULE BUSINESS RULES
# ============================================
catalog_config:
# ------------------------------------------
# PRODUCT LIFECYCLE
# ------------------------------------------
product_lifecycle:
# Default status for newly created products
default_status: "DRAFT"
# Valid status transitions
# DRAFT -> ACTIVE -> DISCONTINUED
# ACTIVE -> DRAFT (revert to draft for editing)
allowed_transitions:
DRAFT: ["ACTIVE"]
ACTIVE: ["DRAFT", "DISCONTINUED"]
DISCONTINUED: ["ACTIVE"] # Re-activate if needed
# Require at least one image before activation
require_image_on_activate: false
# Require barcode before activation
require_barcode_on_activate: true
# Require price before activation
require_price_on_activate: true
# ------------------------------------------
# PRICING
# ------------------------------------------
pricing:
# Price hierarchy levels (highest priority first)
price_hierarchy:
- "price_book_override"
- "channel_price"
- "sale_price"
- "default_price"
- "msrp"
# Require manager approval for markdown
markdown_approval_required: true
# Maximum markdown percentage without owner approval
max_markdown_percent: 50
# Minimum margin threshold (warn if below)
minimum_margin_warning_percent: 10
# Allow selling below cost
allow_below_cost_sale: false
# ------------------------------------------
# BARCODE
# ------------------------------------------
barcode:
# Default barcode format
default_format: "CODE128" # Options: "CODE128", "EAN13", "UPC_A"
# Auto-generate barcode if none provided
auto_generate: true
# Auto-generated barcode prefix (tenant-specific)
auto_prefix: "POS"
# Barcode uniqueness scope
uniqueness_scope: "tenant" # Options: "tenant", "global"
# ------------------------------------------
# CATEGORIES
# ------------------------------------------
categories:
# Maximum category nesting depth
max_depth: 4
# Require every product to have a category
require_category_on_product: true
# Allow product in multiple categories
allow_multi_category: false
# Default category sort order
default_sort: "name_asc"
# ------------------------------------------
# SHOPIFY SYNC
# ------------------------------------------
shopify_sync:
# Scheduled reconciliation interval (minutes)
reconciliation_interval_minutes: 15
# Field ownership model (determines which system can edit each field)
field_ownership:
pos_owned:
- "title"
- "vendor"
- "product_type"
- "variants.price"
- "variants.compare_at_price"
- "variants.sku"
- "variants.barcode"
- "variants.cost"
- "variants.weight"
shopify_owned:
- "body_html"
- "seo_title"
- "seo_description"
- "images"
- "tags"
bidirectional:
- "status"
- "variants.inventory_quantity"
# Conflict resolution for bidirectional fields
conflict_resolution: "pos_wins" # Options: "pos_wins", "shopify_wins", "latest_wins"
# Delete behavior when product discontinued in POS
delete_on_discontinue: false # Set to draft in Shopify, not delete
# Webhook retry on failure
webhook_retry_count: 3
webhook_retry_backoff_seconds: [5, 15, 45]
# ------------------------------------------
# SCANNER CONFIGURATION
# ------------------------------------------
scanner:
# Maximum tags per bulk lookup request
max_tags_per_request: 50
# Tag read timeout (milliseconds)
read_timeout_ms: 5000
# Auto-lookup on scan
auto_lookup_on_scan: true
# Scan confirmation beep
confirmation_beep_enabled: true
Catalog Configuration Field Reference
| Key | Type | Default | Description |
|---|---|---|---|
product_lifecycle.default_status | Enum | DRAFT | Status assigned to newly created products |
product_lifecycle.require_barcode_on_activate | Boolean | true | Barcode must exist before product can transition to ACTIVE |
pricing.markdown_approval_required | Boolean | true | Whether manager must approve markdown price reductions |
pricing.max_markdown_percent | Integer | 50 | Maximum markdown before owner-level approval |
pricing.allow_below_cost_sale | Boolean | false | Whether POS allows sale price below product cost |
barcode.default_format | Enum | CODE128 | Default barcode symbology for auto-generated barcodes |
barcode.auto_generate | Boolean | true | Auto-create barcode if product has no barcode at creation |
categories.max_depth | Integer | 4 | Maximum category tree nesting levels |
categories.require_category_on_product | Boolean | true | Products must be assigned to at least one category |
shopify_sync.reconciliation_interval_minutes | Integer | 15 | Minutes between scheduled full reconciliation |
shopify_sync.conflict_resolution | Enum | pos_wins | Tie-breaking strategy for bidirectional field conflicts |
scanner.max_tags_per_request | Integer | 50 | Maximum scanner tags per bulk API request |
scanner.read_timeout_ms | Integer | 5000 | Milliseconds before scanner read times out |
5.19.4 Inventory Configuration
# ============================================
# INVENTORY MODULE BUSINESS RULES
# ============================================
# All values shown are defaults and can be
# overridden at tenant or store level.
# ============================================
inventory_config:
# ------------------------------------------
# GENERAL SETTINGS
# ------------------------------------------
general:
# Allow inventory to go negative (e.g., sell without stock)
allow_negative_inventory: false
# Costing method for COGS and valuation
costing_method: "weighted_average" # Options: "weighted_average", "fifo"
# Default currency for all inventory valuations
default_currency: "USD"
# Enforce reorder point monitoring
enforce_reorder_points: true
# Minimum display quantity on POS (show "Low Stock" below this)
min_display_qty: 3
# Show exact quantity or generic "In Stock" / "Low Stock"
show_exact_qty_on_pos: true
# ------------------------------------------
# INVENTORY STATUS MODEL
# ------------------------------------------
status_model:
# Valid inventory statuses
statuses:
- "AVAILABLE"
- "RESERVED"
- "QUARANTINE"
- "DAMAGED"
- "IN_TRANSIT"
- "ON_HOLD"
# Only these statuses allow sale
sellable_statuses:
- "AVAILABLE"
# Only these statuses allow transfer out
transferable_statuses:
- "AVAILABLE"
# ------------------------------------------
# PURCHASE ORDERS
# ------------------------------------------
purchase_orders:
# Auto-close PO threshold (% received to auto-complete)
auto_close_threshold_percent: 100
# Allow editing cost at time of receiving
allow_cost_editing_on_receive: true
# Default tax rate applied to PO lines
default_tax_rate: 0.0
# PO approval thresholds (value-based)
approval:
auto_approve_below: 500.00
manager_approval_threshold: 500.00
admin_approval_threshold: 5000.00
approval_expiry_days: 7
# Auto-PO generation
auto_po:
enabled: true
mode: "draft" # Options: "draft", "auto_submit"
minimum_po_value: 50.00
consolidate_by_vendor: true
# PO number format
number_format: "PO-{YEAR}-{SEQUENCE:5}"
# ------------------------------------------
# RECEIVING
# ------------------------------------------
receiving:
# Receiving mode: "open" or "strict"
mode: "open"
# Over-receive threshold (% above PO qty allowed)
over_receive_threshold_percent: 10
# Scanner mode for receiving
scanner_mode: "scan_primary" # Options: "scan_required", "scan_optional", "scan_primary"
# Auto-print labels on receive
auto_print_labels: true
# Label template for received items
label_template: "standard_barcode"
# Discrepancy handling (the "triple approach")
discrepancy:
allow_partial_receive: true
auto_quarantine_damaged: true
auto_create_rma_on_damage: true
cost_variance_alert_percent: 5
# Non-PO receiving
non_po_receiving:
enabled: true
require_reason: true
valid_reasons:
- "CUSTOMER_RETURN"
- "VENDOR_REPLACEMENT"
- "CONSIGNMENT"
- "FOUND_STOCK"
- "OTHER"
# ------------------------------------------
# REORDER & VELOCITY
# ------------------------------------------
reorder:
# Sales velocity calculation window (days)
velocity_window_days: 90
# Safety stock calculation (standard deviations)
safety_stock_sigma: 1.65 # ~95% service level
# Dead stock threshold (days with zero sales)
dead_stock_days: 90
# Seasonal adjustment
seasonal_adjustment_enabled: true
seasonal_history_years: 3
# Reorder point recalculation frequency
recalculation_frequency: "weekly" # Options: "daily", "weekly"
recalculation_day: "sunday"
# Minimum reorder quantity
min_reorder_qty: 1
# Round reorder qty to vendor's case pack size
round_to_case_pack: true
# ------------------------------------------
# COUNTING (STOCKTAKE)
# ------------------------------------------
counting:
# Freeze mode during full physical count
default_freeze_mode: "soft" # Options: "hard", "soft", "none"
# Blind count (hide expected qty from counter)
blind_count_enabled: true
# Scanner mode for counting
scanner_mode: "scan_primary"
# Variance threshold requiring manager approval (units)
variance_approval_threshold_units: 10
# Variance threshold requiring manager approval (%)
variance_approval_threshold_percent: 5
# Auto-adjust variances below threshold
auto_adjust_below_threshold: true
# Require second count for high-variance items
require_recount_above_percent: 20
# Supported count types
count_types:
- "full_physical"
- "cycle_count"
- "spot_check"
- "scanner_assisted"
- "on_demand"
# Cycle count frequency (days between counts per product)
cycle_count_interval_days:
A_items: 30
B_items: 60
C_items: 90
# ------------------------------------------
# ADJUSTMENTS
# ------------------------------------------
adjustments:
# Approval mode for manual adjustments
approval_mode: "threshold" # Options: "all", "threshold", "none"
# Threshold above which approval is required (units)
approval_threshold_units: 10
# Threshold above which approval is required (value)
approval_threshold_value: 100.00
# Default reason codes (built-in)
default_reason_codes:
- code: "DAMAGED"
label: "Damaged / Defective"
requires_note: false
- code: "THEFT"
label: "Theft / Shrinkage"
requires_note: false
- code: "COUNT_CORRECTION"
label: "Count Correction"
requires_note: false
- code: "VENDOR_RETURN"
label: "Returned to Vendor"
requires_note: false
- code: "FOUND_STOCK"
label: "Found Stock (Positive Adjustment)"
requires_note: true
- code: "SAMPLE"
label: "Removed for Sample / Display"
requires_note: false
- code: "DONATION"
label: "Donated"
requires_note: true
- code: "OTHER"
label: "Other"
requires_note: true
# Allow tenants to create custom reason codes
allow_custom_reason_codes: true
# Require note for all adjustments
require_note_for_all: false
# ------------------------------------------
# TRANSFERS
# ------------------------------------------
transfers:
# Require acceptance scan at destination
require_acceptance_at_destination: true
# Allow partial transfer receive
allow_partial_receive: true
# Transfer auto-suggestion
auto_suggest:
enabled: true
imbalance_threshold_days: 14
min_transfer_qty: 2
min_source_qty_after_transfer: 2
frequency: "daily"
# Transfer number format
number_format: "TRF-{YEAR}-{SEQUENCE:5}"
# Transit time estimate (default days)
default_transit_days: 3
# Auto-cancel unfulfilled transfer requests after N days
auto_cancel_unfulfilled_days: 14
# ------------------------------------------
# VENDOR RMA
# ------------------------------------------
rma:
# Allow overstock returns to vendor
overstock_returns_enabled: true
# Default restocking fee (% of cost)
default_restocking_fee_percent: 0
# Maximum restocking fee allowed
max_restocking_fee_percent: 25
# RMA expiry (days to ship back after vendor approval)
ship_back_deadline_days: 30
# Auto-create RMA for damaged items on receive
auto_create_on_damaged_receive: true
# RMA number format
number_format: "RMA-{YEAR}-{SEQUENCE:5}"
# Credit reconciliation reminder (days after shipment)
credit_followup_days: 30
# ------------------------------------------
# SERIAL & LOT TRACKING
# ------------------------------------------
serial_tracking:
# Require serial scan at POS sale
require_serial_at_sale: true
# Require serial scan at receiving
require_serial_at_receive: true
# Serial number format validation (regex, optional)
format_validation: null
lot_tracking:
# FIFO enforcement on sale
fifo_enforcement: true
# Lot number format
number_format: "LOT-{YEAR}{MONTH}-{SEQUENCE:4}"
# Track expiry dates
expiry_tracking_enabled: false # Clothing typically doesn't expire
# ------------------------------------------
# ALERTS & NOTIFICATIONS
# ------------------------------------------
alerts:
low_stock:
enabled: true
delivery: ["dashboard", "email_digest"]
email_digest_time: "07:00"
severity: "warning"
overstock:
enabled: true
delivery: ["dashboard"]
days_of_supply_threshold: 90
severity: "info"
shrinkage:
enabled: true
delivery: ["dashboard", "email_immediate"]
variance_threshold_percent: 5
severity: "critical"
aging_inventory:
enabled: true
delivery: ["dashboard", "email_digest"]
days_threshold: 90
email_digest_day: "monday"
severity: "warning"
po_overdue:
enabled: true
delivery: ["dashboard", "email"]
buffer_days: 3
severity: "warning"
# ------------------------------------------
# OFFLINE INVENTORY
# ------------------------------------------
offline:
# Conflict resolution strategy
conflict_resolution_strategy: "last_write_wins"
# Maximum queued inventory changes
max_queue_size: 500
# Warning threshold (% of queue capacity)
queue_warning_threshold_percent: 90
# Maximum offline duration before forcing reconnect (hours)
max_offline_hours: 24
# Operations allowed offline
allowed_offline:
- "sale_decrement"
- "return_increment"
- "adjustment"
- "count_entry"
- "parked_sale"
# Operations blocked offline
blocked_offline:
- "multi_store_lookup"
- "transfer_request"
- "online_fulfillment"
- "shopify_sync"
- "po_submission"
- "po_receiving"
- "rma_creation"
# ------------------------------------------
# ONLINE FULFILLMENT
# ------------------------------------------
online_fulfillment:
# Store assignment strategy
strategy: "nearest" # Options: "nearest", "round_robin", "priority"
# Allow split fulfillment across multiple stores
split_fulfillment_enabled: false
max_splits: 3
# Exclude HQ warehouse from online fulfillment
exclude_hq: false
# Minimum stock to retain at store after online order reservation
min_retain_qty: 1
# Shopify sync settings
shopify_sync:
reconciliation_interval_minutes: 15
webhook_retry_count: 3
webhook_retry_backoff: [5, 15, 45]
source_of_truth: "pos"
# ------------------------------------------
# POS INTEGRATION
# ------------------------------------------
pos_integration:
# Soft reservation behavior on add to cart
cart_reservation:
enabled: true
type: "soft"
payment_failure_hold_seconds: 30
# Parked sale reservation
parked_sale_reservation:
type: "soft"
show_warning_to_other_terminals: true
# Hold-for-pickup reservation
hold_for_pickup_reservation:
type: "hard"
# Return to stock default status
return_default_status: "AVAILABLE"
# Allow staff to override return status
return_status_override_allowed: true
# Override options
return_status_options:
- "AVAILABLE"
- "DAMAGED"
- "QUARANTINE"
Inventory Configuration Field Reference
| Key | Type | Default | Description |
|---|---|---|---|
general.allow_negative_inventory | Boolean | false | Whether sales can proceed when stock is zero |
general.costing_method | Enum | weighted_average | Valuation method for COGS calculation |
general.min_display_qty | Integer | 3 | POS shows “Low Stock” warning below this threshold |
purchase_orders.approval.auto_approve_below | Decimal | 500.00 | PO total below which no approval is needed |
purchase_orders.approval.admin_approval_threshold | Decimal | 5000.00 | PO total requiring admin/owner approval |
purchase_orders.auto_po.enabled | Boolean | true | Auto-generate draft POs when stock hits reorder point |
receiving.mode | Enum | open | Staff sees expected qty; can receive different amount |
receiving.over_receive_threshold_percent | Integer | 10 | Maximum over-receive without manager approval |
reorder.velocity_window_days | Integer | 90 | Days of sales history for velocity calculation |
reorder.safety_stock_sigma | Decimal | 1.65 | Standard deviations for safety stock (~95% service level) |
reorder.dead_stock_days | Integer | 90 | Days of zero sales before flagging as dead stock |
counting.default_freeze_mode | Enum | soft | Inventory freeze behavior during physical counts |
counting.blind_count_enabled | Boolean | true | Hide expected quantities from counters |
counting.variance_approval_threshold_percent | Integer | 5 | Variance percentage requiring manager approval |
adjustments.approval_mode | Enum | threshold | When manual adjustments require manager approval |
transfers.auto_suggest.imbalance_threshold_days | Integer | 14 | Days-of-supply imbalance to trigger auto-suggestion |
transfers.default_transit_days | Integer | 3 | Default estimated transit time between locations |
rma.ship_back_deadline_days | Integer | 30 | Days to ship RMA items after vendor approval |
offline.max_queue_size | Integer | 500 | Maximum queued offline inventory changes |
offline.max_offline_hours | Integer | 24 | Hours before system forces reconnect attempt |
online_fulfillment.strategy | Enum | nearest | Store selection algorithm for online orders |
online_fulfillment.min_retain_qty | Integer | 1 | Minimum stock to keep at store after online reservation |
pos_integration.cart_reservation.type | Enum | soft | Reservation type when item is added to POS cart |
pos_integration.return_default_status | Enum | AVAILABLE | Default inventory status for returned items |
5.20 Tenant Onboarding Wizard
Scope: Step-by-step guided setup workflow for provisioning a new tenant from initial registration through operational go-live readiness. The onboarding wizard ensures that every required configuration area is addressed before the tenant begins processing transactions.
Cross-Reference: All configuration sections referenced below (5.2 through 5.19) contain the detailed data models and business rules for each step.
5.20.1 Onboarding Flow
The onboarding wizard presents 13 sequential steps. Steps 1-5 are mandatory for go-live. Steps 6-13 are recommended but can be deferred.
flowchart TD
S1["Step 1: Company Info\n(5.2, 5.3)"] --> S2["Step 2: Locations\n(5.4)"]
S2 --> S3["Step 3: Registers\n(5.7)"]
S3 --> S4["Step 4: Printers\n(5.8)"]
S4 --> S5["Step 5: Users & Roles\n(5.5)"]
S5 --> S6["Step 6: Clock-In/Out\n(5.6)"]
S6 --> S7["Step 7: Tax\n(5.9)"]
S7 --> S8["Step 8: Units of Measure\n(5.10)"]
S8 --> S9["Step 9: Payment Methods\n(5.11)"]
S9 --> S10["Step 10: Email\n(5.15)"]
S10 --> S11["Step 11: Integrations\n(5.16)"]
S11 --> S12["Step 12: Business Rules\n(5.19)"]
S12 --> S13["Step 13: Go-Live Checklist\n(Validation)"]
S13 -->|All checks pass| GL["GO LIVE"]
S13 -->|Checks failed| FIX["Review Failures\n(Return to failing step)"]
FIX --> S13
style S1 fill:#2d6a4f,stroke:#1b4332,color:#fff
style S2 fill:#2d6a4f,stroke:#1b4332,color:#fff
style S3 fill:#2d6a4f,stroke:#1b4332,color:#fff
style S4 fill:#264653,stroke:#1d3557,color:#fff
style S5 fill:#2d6a4f,stroke:#1b4332,color:#fff
style S6 fill:#264653,stroke:#1d3557,color:#fff
style S7 fill:#2d6a4f,stroke:#1b4332,color:#fff
style S8 fill:#264653,stroke:#1d3557,color:#fff
style S9 fill:#2d6a4f,stroke:#1b4332,color:#fff
style S10 fill:#264653,stroke:#1d3557,color:#fff
style S11 fill:#264653,stroke:#1d3557,color:#fff
style S12 fill:#264653,stroke:#1d3557,color:#fff
style S13 fill:#7b2d8e,stroke:#5a1d6e,color:#fff
style GL fill:#2d6a4f,stroke:#1b4332,color:#fff
style FIX fill:#c0392b,stroke:#922b21,color:#fff
5.20.2 Step Details
| Step | Name | Section Ref | Required | Description |
|---|---|---|---|---|
| 1 | Company Info | 5.2, 5.3 | Yes | Set tenant name, legal entity name, logo, timezone, base currency, date/time format, fiscal year start |
| 2 | Locations | 5.4 | Yes | Create all physical locations (stores and warehouses). Set address, phone, timezone override, and location type. |
| 3 | Registers | 5.7 | Yes | Add registers to each store location. Select profile (Full POS, Mobile POS, Inventory-Only). Pair physical devices. |
| 4 | Printers | 5.8 | No | Register receipt and label printers. Link printers to registers. Run network discovery if applicable. |
| 5 | Users & Roles | 5.5 | Yes | Create user accounts. Assign roles (Owner, Admin, Manager, Cashier, Inventory Clerk). Assign users to locations. |
| 6 | Clock-In/Out | 5.6 | No | Configure clock-in/clock-out settings. Enable time tracking for staff if needed. Skip for operations that do not require time tracking. |
| 7 | Tax | 5.9 | Yes | Assign tax jurisdiction to each store location. Review compound tax rates (State/County/City), tax calculation priority, and exemption handling. |
| 8 | Units of Measure | 5.10 | No | Review predefined UoMs. Create custom UoMs if needed (e.g., BUNDLE, SET, ROLL). Skip if standard UoMs are sufficient. |
| 9 | Payment Methods | 5.11 | Yes | Enable payment methods per location (Cash, Credit Card, Gift Card, Store Credit, etc.). Configure payment processor credentials. |
| 10 | 5.15 | No | Configure SMTP or API email provider. Send test email. Enable/disable individual email templates. | |
| 11 | Integrations | 5.16 | No | Connect Shopify store (enter shop URL, API key, access token). Verify connection. Configure sync mode. |
| 12 | Business Rules | 5.19 | No | Review all business rule defaults. Customize return policy, discount limits, offline mode settings, and inventory rules as needed. |
| 13 | Go-Live Checklist | – | Yes | Automated validation of all mandatory requirements. Displays pass/fail for each check. |
5.20.3 Go-Live Validation Rules
The go-live checklist runs automated validation against all mandatory configuration requirements. The tenant cannot begin processing transactions until all mandatory checks pass.
Mandatory Checks (must all pass):
| # | Validation Rule | Failure Message |
|---|---|---|
| 1 | At least 1 location created | “No locations configured. Create at least one store location.” |
| 2 | At least 1 location of type STORE | “No store locations found. At least one location must be a retail store.” |
| 3 | At least 1 register per store location | “Store ‘{location_name}’ has no registers. Add at least one register.” |
| 4 | At least 1 user with OWNER role | “No Owner user found. At least one user must have the Owner role.” |
| 5 | Tax jurisdiction assigned to each store location | “Store ‘{location_name}’ has no tax jurisdiction assigned. Assign a tax jurisdiction with at least one active rate.” |
| 6 | At least 1 payment method enabled per store location | “Store ‘{location_name}’ has no payment methods enabled. Enable at least one.” |
| 7 | Email sender configured | “No email provider configured. Configure SMTP or API email for receipts.” |
| 8 | Base currency set | “Base currency is not configured. Set the tenant’s base currency.” |
| 9 | Default timezone set | “Default timezone is not configured. Set the tenant’s timezone.” |
Recommended Checks (advisory, do not block go-live):
| # | Validation Rule | Advisory Message |
|---|---|---|
| 1 | Shopify integration connected | “Shopify integration is not connected. Online orders will not sync.” |
| 2 | At least 1 label printer per location | “No label printers configured. Barcode printing will be unavailable.” |
| 3 | Receipt printer linked to each Full POS register | “Register ‘{register_name}’ has no receipt printer linked.” |
| 4 | Custom fields configured (if applicable) | “No custom fields defined. Custom fields can be added later.” |
| 5 | Business rules reviewed | “Business rules are using system defaults. Review and customize as needed.” |
| 6 | Clock-in/out reviewed | “Clock-in/out settings not reviewed. Time tracking will use system defaults.” |
| 7 | Receipt header/footer customized | “Receipt template is using system defaults. Customize header and footer.” |
5.20.4 Onboarding State Tracking
Onboarding Progress Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant (unique constraint) |
current_step | Integer | Yes | Current wizard step (1-14) |
steps_completed | JSON | Yes | Map of step_number: {completed: boolean, completed_at: DateTime, skipped: boolean} |
go_live_ready | Boolean | Yes | Whether all mandatory checks pass (default: false) |
go_live_at | DateTime | No | Timestamp when the tenant was activated for live operations |
started_at | DateTime | Yes | Timestamp when onboarding began |
completed_at | DateTime | No | Timestamp when the wizard was fully completed (all 14 steps addressed) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Steps can be completed in any order, though the wizard presents them sequentially as a recommended progression.
- Steps can be revisited and modified at any time before go-live.
- Steps marked as “No” in the Required column can be explicitly skipped. Skipping sets
skipped: truefor that step. - After go-live, the onboarding wizard is no longer displayed. All configuration is managed through the standard Admin Portal settings pages.
- The onboarding wizard state is preserved so that if the tenant administrator abandons the wizard mid-way and returns later, progress is restored.
5.21 User Stories & Acceptance Criteria
Scope: All user stories and Gherkin acceptance criteria for Module 5 (Setup & Configuration). Stories are organized into 17 epics covering all functional areas. Each epic includes user stories in standard format and one or more Gherkin feature files with acceptance scenarios.
Epic 5.A: System Settings & Branding
US-5.A.1: Configure Company Identity
- As a tenant administrator, I want to configure the company name, legal entity name, logo, and timezone so that the system reflects our brand identity across all interfaces and reports.
- Constraint: Logo must be PNG or SVG, max 2MB, min 200x200px. Timezone uses IANA identifiers.
US-5.A.2: Set Business Hours per Location
- As a tenant administrator, I want to set operating hours per location so that shift schedules and report periods align with actual store hours.
- Constraint: Hours are defined per day of week. Locations can have different hours.
US-5.A.3: Customize Branding
- As a tenant administrator, I want to customize the login page branding, primary color scheme, and admin portal header so that the system looks consistent with our brand.
- Constraint: Primary and secondary color must be valid hex codes. Preview available before saving.
Feature: System Settings Configuration
As a tenant administrator
I need to configure company identity and branding
So that the system reflects our business across all touchpoints
Background:
Given I am logged in as a user with "OWNER" role
And I am on the System Settings page
Scenario: Configure company identity on initial setup
When I enter "Nexus Clothing" as the tenant name
And I enter "Nexus Clothing LLC" as the legal entity name
And I upload a valid PNG logo "nexus-logo.png" (500x500px, 150KB)
And I select timezone "America/New_York"
And I select currency "USD"
And I click "Save Settings"
Then the settings should be saved successfully
And the header should display "Nexus Clothing"
And the logo should appear in the top-left corner
Scenario: Reject invalid logo upload
When I attempt to upload a logo file "oversized.png" (5MB)
Then the system should reject the upload with message "Logo must be under 2MB"
And the existing logo should remain unchanged
Scenario: Currency becomes immutable after first transaction
Given at least one transaction has been processed
When I attempt to change the base currency from "USD" to "CAD"
Then the system should block the change with message "Currency cannot be changed after transactions have been processed"
Epic 5.B: Location Management
US-5.B.1: Create Store and Warehouse Locations
- As a tenant administrator, I want to create store and warehouse locations with addresses and contact information so that the system knows the physical topology of the business.
- Constraint: Each location must have a unique name within the tenant. Location type is
STOREorWAREHOUSE.
Feature: Location Management
As a tenant administrator
I need to define physical locations
So that the system maps to our real-world store topology
Background:
Given I am logged in as a user with "ADMIN" role
Scenario: Create a new store location
When I navigate to Locations management
And I click "Add Location"
And I enter name "Georgetown Store"
And I select type "STORE"
And I enter address "3100 M Street NW, Washington, DC 20007"
And I enter phone "(202) 555-0100"
And I click "Save Location"
Then the location "Georgetown Store" should be created
Scenario: Prevent duplicate location names
Given a location named "Georgetown Store" already exists
When I attempt to create another location named "Georgetown Store"
Then the system should reject with message "A location with this name already exists"
Epic 5.C: User & Role Management
US-5.C.1: Create User Accounts
- As a tenant administrator, I want to create user accounts with profiles including name, email, phone, and PIN so that staff can access the system with appropriate credentials.
- Constraint: Email must be unique within the tenant. PIN must be exactly 4 digits. Username auto-generated from email.
US-5.C.2: Configure Feature Toggles per Role
- As a tenant administrator, I want to configure which features each role can access so that staff only see functionality relevant to their job.
- Constraint: Feature toggles are grouped by module. Changes take effect on next login.
US-5.C.3: Assign Users to Multiple Locations
- As a tenant administrator, I want to assign a user to one or more locations so that I can set their default location, control reporting filters, and organize staff by store.
- Constraint: A user must be assigned to at least one location. Location assignments are informational — they set defaults and reporting scope but do not restrict which locations a user can process transactions at.
Feature: User and Role Management
As a tenant administrator
I need to manage staff accounts, roles, and location assignments
So that employees have appropriate defaults and reporting scope
Background:
Given I am logged in as a user with "OWNER" role
And locations "Georgetown Store" and "HQ Warehouse" exist
Scenario: Create a new cashier user
When I navigate to User Management
And I click "Add User"
And I enter first name "Sarah" and last name "Johnson"
And I enter email "sarah.johnson@nexus.com"
And I enter PIN "4521"
And I assign role "CASHIER"
And I assign location "Georgetown Store"
And I click "Create User"
Then user "Sarah Johnson" should be created with role "CASHIER"
And she should be assigned to "Georgetown Store"
And a welcome email should be queued
Scenario: Prevent duplicate email
Given user "sarah.johnson@nexus.com" already exists
When I attempt to create a user with email "sarah.johnson@nexus.com"
Then the system should reject with message "A user with this email already exists"
Scenario: Assign user to multiple locations (informational)
Given user "Mark Chen" exists with role "MANAGER"
When I edit user "Mark Chen"
And I add location assignment "HQ Warehouse"
And I click "Save"
Then "Mark Chen" should be assigned to both "Georgetown Store" and "HQ Warehouse"
And his primary location should be used as the default at login
And he should be able to process transactions at any tenant location regardless of assignment
Epic 5.D: Register & Device Management
US-5.D.1: Add Registers to a Location
- As a tenant administrator, I want to add registers to a store location so that the store can process transactions.
- Constraint: Register name must be unique within the location. Register number auto-increments.
US-5.D.2: Pair Devices with Registers
- As a tenant administrator, I want to pair a physical device (iPad, PC, terminal) with a register so that the device operates as that register.
- Constraint: Device pairing uses a one-time PIN code. One device per register at a time.
US-5.D.3: Assign Register Profiles
- As a tenant administrator, I want to assign a profile (Full POS, Mobile POS, Inventory-Only) to each register so that the interface matches the register’s purpose.
- Constraint: Full POS requires a receipt printer link. Mobile POS and Inventory-Only do not require printers.
Feature: Register and Device Management
As a tenant administrator
I need to configure registers and pair physical devices
So that store staff can process transactions
Background:
Given I am logged in as a user with "ADMIN" role
And location "Georgetown Store" exists
Scenario: Add a Full POS register
When I navigate to Registers for "Georgetown Store"
And I click "Add Register"
And I enter name "Main Counter"
And I select profile "FULL_POS"
And I click "Save Register"
Then register "Main Counter" should be created at "Georgetown Store"
And a configuration warning should show "No receipt printer linked"
And the register status should be "INACTIVE" until a device is paired
Scenario: Pair a device using PIN code
Given register "Main Counter" exists at "Georgetown Store"
When I click "Generate Pairing PIN" on the register
Then a 6-digit pairing PIN should be displayed with 5-minute expiry
When the POS device enters the pairing PIN "847293"
Then the device should be paired to "Main Counter"
And the register status should change to "ACTIVE"
Scenario: Prevent dual device pairing
Given register "Main Counter" is already paired to "iPad-001"
When I attempt to pair "iPad-002" to "Main Counter"
Then the system should prompt "This register is already paired to iPad-001. Unpair first?"
Epic 5.E: Printer Configuration
US-5.E.1: Register Printers
- As a tenant administrator, I want to register receipt and label printers at each location so that transactions can print receipts and staff can print barcode labels.
- Constraint: Printers are classified as RECEIPT or LABEL. Connection types: USB, NETWORK_IP, BLUETOOTH.
US-5.E.2: Link Printers to Registers
- As a tenant administrator, I want to link printers to specific registers so that each register knows which printer to use.
- Constraint: Each Full POS register requires exactly one PRIMARY_RECEIPT printer. LABEL and SECONDARY_RECEIPT are optional.
Feature: Printer Configuration
As a tenant administrator
I need to register and link printers to registers
So that receipts and labels print to the correct devices
Background:
Given I am logged in as a user with "ADMIN" role
And location "Georgetown Store" exists with register "Main Counter"
Scenario: Register a network receipt printer
When I navigate to Printers for "Georgetown Store"
And I click "Add Printer"
And I enter name "Main Counter Printer"
And I select type "RECEIPT"
And I select connection "NETWORK_IP"
And I enter address "192.168.1.50:9100"
And I select paper width "80MM"
And I click "Save Printer"
Then printer "Main Counter Printer" should be registered
And a health check should be initiated
And status should display "ONLINE" or "OFFLINE"
Scenario: Link receipt printer to register
Given printer "Main Counter Printer" exists and is ONLINE
When I navigate to register "Main Counter" printer assignments
And I assign "Main Counter Printer" as "PRIMARY_RECEIPT"
And I click "Save"
Then the configuration warning "No receipt printer linked" should be cleared
And "Main Counter" should print to "Main Counter Printer"
Scenario: USB printer cannot be shared
Given a USB printer "USB Receipt Printer" is linked to "Main Counter"
When I attempt to link "USB Receipt Printer" to register "Side Counter"
Then the system should reject with message "USB printers cannot be shared between registers"
Epic 5.F: Tax Configuration
US-5.F.1: Assign Tax Jurisdiction to Location
- As a tenant administrator, I want to assign a tax jurisdiction to each store location so that compound sales tax (State + County + City) is correctly calculated on transactions.
- Constraint: Each location references exactly one jurisdiction. A jurisdiction defines up to 3 rate levels. Rate changes per level are scheduled with an effective date.
US-5.F.2: Automatic Compound Tax Calculation
- As a cashier, I want tax to automatically calculate on each line item by summing all active rates for the store location’s jurisdiction so that I do not need to manually compute tax.
- Constraint: Tax follows priority: product exemption > customer exemption > jurisdiction compound rate.
Feature: Tax Jurisdiction Configuration
As a tenant administrator
I need to configure tax jurisdictions and assign them to store locations
So that compound sales tax is correctly applied to all transactions
Background:
Given I am logged in as a user with "ADMIN" role
And location "Georgetown Store" exists in Virginia
Scenario: Create a tax jurisdiction and assign to location
When I navigate to Tax Jurisdictions
And I create jurisdiction "VA-NFK" named "Norfolk, Virginia"
And I add a STATE rate "Virginia State Tax" at 4.300%
And I add a COUNTY rate "Hampton Roads Regional Tax" at 0.700%
And I add a CITY rate "Norfolk City Tax" at 1.000%
And I assign jurisdiction "VA-NFK" to "Georgetown Store"
Then the effective compound tax rate for "Georgetown Store" should be 6.000%
And all future transactions at this location should apply 6.000% compound tax
Scenario: Schedule a future rate change at one level
Given "Georgetown Store" has jurisdiction "VA-NFK" with compound rate 6.000%
When I add a new STATE rate of "4.500%" with effective date "2027-01-01"
And I click "Save Tax Rate"
Then the system should show the current compound rate as 6.000%
And the scheduled STATE rate as 4.500% effective January 1, 2027
And the compound rate should automatically change to 6.200% on the effective date
Scenario: Tax exemption applied for exempt customer
Given "Georgetown Store" has a compound tax rate of 6.000%
And customer "City of Alexandria" is marked as tax-exempt with valid certificate
When a sale is processed for "City of Alexandria"
Then all line items should show tax as $0.00
And the receipt should display "Tax Exempt" with the certificate number
Epic 5.G: UoM Management
US-5.G.1: Manage Units of Measure
- As a tenant administrator, I want to view and manage the available units of measure so that products use the correct selling and purchasing units.
- Constraint: System-predefined UoMs cannot be modified or deleted. Custom UoMs can be deactivated.
US-5.G.2: Create Custom UoMs
- As a tenant administrator, I want to create custom units of measure with conversion factors so that I can handle specialty product measurements.
- Constraint: Custom UoMs must specify a category (QUANTITY, LENGTH, WEIGHT) and conversion factor to the base unit.
Feature: Unit of Measure Management
As a tenant administrator
I need to manage units of measure
So that products are sold and purchased in the correct units
Background:
Given I am logged in as a user with "ADMIN" role
Scenario: View predefined UoMs
When I navigate to Units of Measure management
Then I should see system UoMs including "Each", "Pair", "Dozen", "Case", "Yard"
And system UoMs should be marked as non-editable
Scenario: Create a custom UoM
When I click "Add Custom UoM"
And I enter code "BUNDLE" and name "Bundle of 5"
And I select category "QUANTITY"
And I enter conversion factor 5 to base unit "Each"
And I click "Save"
Then UoM "BUNDLE" should be created
And the conversion table should show "1 BUNDLE = 5 EACH"
And the inverse "1 EACH = 0.2 BUNDLE" should be auto-generated
Scenario: Prevent deleting a UoM in use
Given UoM "BUNDLE" is assigned to product "Sock Pack"
When I attempt to delete UoM "BUNDLE"
Then the system should reject with message "Cannot delete UoM in use. Deactivate instead."
Epic 5.H: Payment Methods Setup
US-5.H.1: Enable Payment Methods per Location
- As a tenant administrator, I want to enable or disable payment methods per location so that each store only accepts configured payment types.
- Constraint: At least one payment method must be enabled per store location. Cash cannot be disabled.
US-5.H.2: Configure Payment Processor
- As a tenant administrator, I want to configure the payment processor credentials so that credit card transactions are processed.
- Constraint: Credentials are encrypted at rest. Test transaction available for validation.
Feature: Payment Methods Setup
As a tenant administrator
I need to configure payment methods per location
So that each store can accept the correct payment types
Background:
Given I am logged in as a user with "ADMIN" role
And location "Georgetown Store" exists
Scenario: Enable standard payment methods
When I navigate to Payment Methods for "Georgetown Store"
And I enable "Cash", "Credit Card", "Gift Card", and "Store Credit"
And I disable "Affirm"
And I click "Save"
Then "Georgetown Store" should accept Cash, Credit Card, Gift Card, and Store Credit
And Affirm should not be available as a payment option at this location
Scenario: Prevent disabling all payment methods
When I attempt to disable all payment methods for "Georgetown Store"
Then the system should reject with message "At least one payment method must be enabled"
Scenario: Configure payment processor credentials
When I navigate to Payment Processor settings
And I enter merchant ID "MID_12345"
And I enter API key and secret
And I click "Test Connection"
Then the system should perform a $0.00 authorization test
And the result should display "Connection Successful" or an error message
Epic 5.I: Custom Fields
US-5.I.1: Define Custom Fields per Entity
- As a tenant administrator, I want to define custom fields for products, customers, and transactions so that I can capture business-specific data not covered by standard fields.
- Constraint: Max 20 custom fields per entity type. Supported types: TEXT, NUMBER, DATE, BOOLEAN, SELECT.
US-5.I.2: Custom Fields on Forms
- As a staff member, I want custom fields to appear on the relevant entity forms so that I can enter the custom data during normal workflows.
- Constraint: Custom fields appear in a dedicated “Custom Fields” section. Sort order controls display sequence.
Feature: Custom Fields Configuration
As a tenant administrator
I need to define custom fields for business entities
So that we can capture data specific to our business
Background:
Given I am logged in as a user with "ADMIN" role
Scenario: Create a custom text field for products
When I navigate to Custom Fields management
And I select entity type "PRODUCT"
And I click "Add Custom Field"
And I enter label "Fabric Composition" and field type "TEXT"
And I set max length to 200
And I mark it as required
And I click "Save"
Then custom field "Fabric Composition" should be added to products
And it should appear on the product edit form
Scenario: Create a select field with options
When I add a custom field for "CUSTOMER"
And I enter label "Preferred Contact Method" and field type "SELECT"
And I add options: "Phone", "Email", "SMS", "None"
And I click "Save"
Then the field should appear as a dropdown on customer forms
And only the defined options should be selectable
Scenario: Enforce custom field limit
Given 20 custom fields already exist for entity type "PRODUCT"
When I attempt to add a 21st custom field
Then the system should reject with message "Maximum of 20 custom fields per entity type"
Epic 5.J: Approval Workflows
US-5.J.1: Configure Approval Rules
- As a tenant administrator, I want to configure which actions require approval and the threshold values so that high-value or sensitive operations are reviewed before proceeding.
- Constraint: Approval rules support threshold-based, always-required, and never-required modes per action type.
US-5.J.2: Approve or Reject Requests
- As a manager, I want to view pending approval requests and approve or reject them with an optional reason so that the workflow continues without delay.
- Constraint: A user cannot approve their own request. Pending requests expire after 90 days.
Feature: Approval Workflows
As a tenant administrator
I need to configure approval rules for sensitive actions
So that high-value operations receive proper authorization
Background:
Given I am logged in as a user with "OWNER" role
And approval rules are configured with defaults
Scenario: Configure PO approval threshold
When I navigate to Approval Rules
And I set "PURCHASE_ORDER" manager threshold to $1000.00
And I set "PURCHASE_ORDER" admin threshold to $10000.00
And I click "Save"
Then POs below $1000 should auto-approve
And POs between $1000 and $9999.99 should require manager approval
And POs at $10000+ should require admin approval
Scenario: Approve a pending PO request
Given a PO "PO-2026-00200" for $2500.00 is pending manager approval
And I am logged in as a "MANAGER"
When I navigate to Pending Approvals
And I select "PO-2026-00200"
And I click "Approve"
Then the PO status should change to "APPROVED"
And the requester should receive a notification
Scenario: Self-approval is prevented
Given user "John" submitted PO "PO-2026-00201"
When "John" attempts to approve his own PO
Then the system should reject with message "You cannot approve your own request"
Epic 5.K: Receipt Builder
US-5.K.1: Configure Receipt Layout
- As a tenant administrator, I want to configure the receipt paper width, font size, field order, and separator style so that receipts match our brand and printer capabilities.
- Constraint: Preview must be available before saving. Changes apply to all registers at the affected location.
US-5.K.2: Customize Header and Footer
- As a tenant administrator, I want to customize the receipt header (company name, tagline, logo) and footer (return policy, website, thank-you message) so that receipts include our business information.
- Constraint: Header supports 3 text lines plus logo. Footer supports 3 text lines. Blank lines are omitted from print.
Feature: Receipt Builder
As a tenant administrator
I need to configure receipt layout and content
So that printed receipts match our brand and include required information
Background:
Given I am logged in as a user with "ADMIN" role
Scenario: Configure receipt header
When I navigate to Receipt Configuration
And I set header line 1 to "Nexus Clothing"
And I set header line 2 to "Fashion for Everyone"
And I upload a monochrome logo
And I set footer line 1 to "Returns accepted within 30 days with receipt."
And I set footer line 2 to "www.nexusclothing.com"
And I click "Save"
Then the receipt preview should show the updated header and footer
And all future receipts should use this configuration
Scenario: Toggle receipt fields
When I navigate to Receipt Field Configuration
And I disable "register_number" and "savings_total"
And I enable "loyalty_points" and "customer_name"
And I click "Save"
Then receipts should show loyalty points and customer name
And receipts should not show register number or savings total
Scenario: Location-level receipt override
Given a tenant-wide receipt configuration exists
When I create a location-specific override for "Georgetown Store"
And I set header line 2 to "Georgetown Location - Since 2015"
And I click "Save"
Then "Georgetown Store" should use the override header
And all other locations should use the tenant-wide default
Epic 5.L: Email & Communications
US-5.L.1: Configure Email Provider
- As a tenant administrator, I want to configure the SMTP or API email provider so that the system can send transactional and notification emails.
- Constraint: Configuration must be verified via test email before activation.
US-5.L.2: Enable or Disable Templates
- As a tenant administrator, I want to enable or disable individual email templates so that I control which automated emails are sent.
- Constraint: Disabling a template prevents sending but does not affect the triggering event.
Feature: Email Configuration
As a tenant administrator
I need to configure email delivery and manage templates
So that customers and staff receive appropriate automated communications
Background:
Given I am logged in as a user with "ADMIN" role
Scenario: Configure SMTP email provider
When I navigate to Email Configuration
And I select provider type "SMTP"
And I enter host "smtp.gmail.com" and port 587
And I enter username "pos@nexusclothing.com"
And I enter password and sender name "Nexus Clothing"
And I click "Send Test Email"
Then a test email should be sent to "pos@nexusclothing.com"
And the configuration should be marked as "Verified"
Scenario: Disable a template
Given all email templates are enabled by default
When I navigate to Email Templates
And I disable template "TMPL-TIER-UPGRADE"
And I click "Save"
Then when a customer's tier changes, no email should be sent
And the tier upgrade should still be applied normally
Scenario: Unverified configuration shows warning
Given no email provider has been configured
When I navigate to the Admin Portal dashboard
Then a warning banner should display "Email provider not configured -- email receipts and notifications will not be sent"
Epic 5.M: Integration Hub
US-5.M.1: Connect Shopify Integration
- As a tenant administrator, I want to connect my Shopify store by entering API credentials so that products and inventory sync between the POS and online store.
- Constraint: Connection verified via test API call. Webhook endpoints auto-registered on successful connection.
US-5.M.2: View Integration Health
- As a tenant administrator, I want to view a dashboard showing the health status of all integrations so that I can identify and troubleshoot sync issues.
- Constraint: Dashboard shows status, last sync time, error count, and latency per integration.
Feature: Integration Hub
As a tenant administrator
I need to connect and monitor external system integrations
So that data flows correctly between the POS and external platforms
Background:
Given I am logged in as a user with "ADMIN" role
Scenario: Connect Shopify store
When I navigate to Integration Hub
And I click "Connect Shopify"
And I enter shop URL "nexus-clothes.myshopify.com"
And I enter API key, API secret, and access token
And I select sync mode "pos_master"
And I click "Verify Connection"
Then the system should make a test API call to Shopify
And the Shopify integration status should change to "CONNECTED"
And webhook endpoints should be auto-registered
Scenario: Integration health alert on errors
Given the Shopify integration has accumulated 6 errors in the past 24 hours
Then the Integration Hub dashboard should show Shopify health as "Red"
And a dashboard alert should be sent to all ADMIN and OWNER users
And the error log should be accessible from the integration detail page
Scenario: Disable integration temporarily
Given Shopify integration is currently "CONNECTED" and "ENABLED"
When I toggle the integration to "DISABLED"
Then all sync operations should halt immediately
And queued webhooks should be preserved
And the status should show "CONNECTED" but "DISABLED"
Epic 5.N: Loyalty Settings
US-5.N.1: Configure Loyalty Rates and Tiers
- As a tenant administrator, I want to configure the points-per-dollar rate, tier thresholds, and tier multipliers so that the loyalty program matches our business strategy.
- Constraint: Tier thresholds must be in ascending order. Point multiplier must be >= 1.0.
US-5.N.2: Configure Gift Card Settings
- As a tenant administrator, I want to configure gift card denominations, load limits, and expiry rules so that gift cards comply with our jurisdiction requirements.
- Constraint: Minimum load must be >= $1.00. Maximum load must be >= minimum load. Expiry must comply with state law.
Feature: Loyalty and Rewards Settings
As a tenant administrator
I need to configure loyalty program parameters
So that the rewards program operates according to our business rules
Background:
Given I am logged in as a user with "OWNER" role
Scenario: Configure loyalty point rates
When I navigate to Loyalty Settings
And I set points per dollar to 2
And I set redemption rate to 200 points per dollar
And I set minimum redemption to 200 points
And I set max redemption percent to 25
And I click "Save"
Then customers should earn 2 base points per dollar
And 200 points should be required for $1.00 discount
And customers can pay up to 25% of their total with points
Scenario: Configure tier thresholds
When I set Silver threshold to $500 with 1.25x multiplier
And I set Gold threshold to $2500 with 1.75x multiplier
And I set Platinum threshold to $7500 with 2.5x multiplier
And I click "Save"
Then the tier thresholds should be updated
And existing customers should be re-evaluated against new thresholds
Scenario: Validate tier threshold ordering
When I attempt to set Silver threshold to $5000 and Gold threshold to $2000
Then the system should reject with message "Tier thresholds must be in ascending order"
Epic 5.O: Audit Configuration
US-5.O.1: Toggle Audit Categories
- As a tenant administrator, I want to enable or disable specific audit log categories so that I can control the volume and focus of audit logging.
- Constraint: At least LOGIN and VOID categories must remain enabled. Other categories can be toggled freely.
US-5.O.2: Set Audit Retention Period
- As a tenant administrator, I want to set the audit log retention period and archival policy so that logs are retained long enough for compliance but do not consume unlimited storage.
- Constraint: Minimum retention is 90 days. Minimum archive retention is 365 days.
Feature: Audit Configuration
As a tenant administrator
I need to configure audit logging categories and retention
So that audit data is captured appropriately and stored compliantly
Background:
Given I am logged in as a user with "OWNER" role
Scenario: Disable a non-mandatory audit category
When I navigate to Audit Configuration
And I disable category "DISCOUNT"
And I click "Save"
Then discount events should no longer generate audit log entries
And existing discount audit entries should remain unchanged
Scenario: Prevent disabling mandatory categories
When I attempt to disable category "LOGIN"
Then the system should reject with message "LOGIN audit category cannot be disabled"
When I attempt to disable category "VOID"
Then the system should reject with message "VOID audit category cannot be disabled"
Scenario: Configure retention and archival
When I set retention period to 180 days
And I enable archival with format "COMPRESSED_JSON"
And I set archive purge to 2555 days (7 years)
And I click "Save"
Then logs older than 180 days should be archived during the next nightly job
And archived logs older than 7 years should be permanently deleted
Epic 5.P: Business Rules Configuration
US-5.P.1: Review and Modify Defaults
- As a tenant administrator, I want to review all business rule defaults and modify values that do not fit our operations so that the system behavior matches our policies.
- Constraint: Changes are validated against business rule constraints (e.g., max_line_discount_percent must be 0-100). Changes take effect immediately.
US-5.P.2: Override Rules at Store Level
- As a tenant administrator, I want to override specific business rules at the store level so that different locations can have different policies where needed.
- Constraint: Store-level overrides take precedence over tenant-level defaults. Only overridden values differ; all others inherit from tenant level.
Feature: Business Rules Configuration
As a tenant administrator
I need to review and customize business rules
So that the system enforces our specific operational policies
Background:
Given I am logged in as a user with "OWNER" role
Scenario: Modify return policy
When I navigate to Business Rules > Sales Configuration
And I change "full_refund_days" from 30 to 14
And I change "store_credit_days" from 90 to 60
And I click "Save"
Then the return policy should update immediately
And new transactions should use 14-day refund and 60-day store credit policies
Scenario: Store-level override for discount limits
When I navigate to Business Rules for "Georgetown Store"
And I override "max_line_discount_percent" to 25 (tenant default is 20)
And I click "Save"
Then "Georgetown Store" should allow up to 25% line discounts
And all other stores should continue using the 20% tenant default
Scenario: Validate business rule constraints
When I attempt to set "max_line_discount_percent" to 150
Then the system should reject with message "Value must be between 0 and 100"
When I attempt to set "cash_drawer.variance_tolerance" to -5.00
Then the system should reject with message "Value must be greater than or equal to 0"
Epic 5.Q: Tenant Onboarding
US-5.Q.1: Provision New Tenant via Wizard
- As a platform administrator, I want to provision a new tenant using the step-by-step onboarding wizard so that all required configuration is completed before the tenant goes live.
- Constraint: Wizard tracks progress across sessions. Steps can be revisited. Mandatory steps cannot be skipped.
US-5.Q.2: Validate Go-Live Readiness
- As a platform administrator, I want the system to validate all go-live requirements and display a pass/fail checklist so that I can confirm the tenant is ready for production operations.
- Constraint: All 9 mandatory checks must pass. Advisory checks are informational only. Go-live timestamp is recorded permanently.
Feature: Tenant Onboarding Wizard
As a platform administrator
I need to provision and validate new tenants
So that they are fully configured before processing transactions
Background:
Given I am provisioning a new tenant "Nexus Clothing"
And I am on the onboarding wizard
Scenario: Complete mandatory onboarding steps
When I complete Step 1 (Company Info) with name "Nexus Clothing", timezone "America/New_York", currency "USD"
And I complete Step 2 (Locations) by creating "Georgetown Store" (type: STORE)
And I complete Step 4 (Registers) by adding "Main Counter" (profile: FULL_POS) to "Georgetown Store"
And I complete Step 6 (Users) by creating user "Will" with role "OWNER" at "Georgetown Store"
And I complete Step 8 (Tax) by setting rate 6.000% for "Georgetown Store"
And I complete Step 10 (Payment Methods) by enabling "Cash" and "Credit Card" at "Georgetown Store"
And I complete Step 11 (Email) by configuring SMTP provider
Then the wizard should show steps 1, 2, 4, 6, 8, 10, and 11 as completed
Scenario: Go-live validation passes with all mandatory requirements
Given all mandatory onboarding steps are completed
When I navigate to Step 14 (Go-Live Checklist)
Then all 9 mandatory checks should show "PASS"
And the "Go Live" button should be enabled
When I click "Go Live"
Then the tenant status should change to "ACTIVE"
And the go-live timestamp should be recorded
And the onboarding wizard should no longer be displayed
Scenario: Go-live validation fails with missing requirements
Given Step 7 (Tax) has not been completed for "Georgetown Store"
When I navigate to Step 13 (Go-Live Checklist)
Then mandatory check "Tax jurisdiction assigned to each store location" should show "FAIL"
And the failure message should read "Store 'Georgetown Store' has no tax jurisdiction assigned. Assign a tax jurisdiction with at least one active rate."
And the "Go Live" button should be disabled
And a link should navigate to Step 8 (Tax Configuration)
End of Module 5: Setup & Configuration (Sections 5.15 – 5.21)
6. Integrations & External Systems Module
6.1 Overview & Scope
6.1.1 Module Purpose
Module 6 consolidates every external-system integration into a single, dedicated module. Rather than scattering API credentials, retry logic, webhook handlers, and protocol-specific code across the Sales, Catalog, Inventory, and Setup modules, the Integration Module provides a unified abstraction layer that all other modules call when they need to communicate with the outside world.
This design delivers three key benefits:
- Single Responsibility – Each business module owns its domain logic; Module 6 owns the wire protocol, authentication, error handling, and data mapping for every external provider.
- Swap-ability – Adding a new payment processor or replacing an email provider requires changes only inside Module 6. Upstream modules remain untouched.
- Operational Visibility – Centralised logging, circuit breakers, rate-limit tracking, and webhook ingestion give the operations team one place to monitor every integration.
6.1.2 Scope Statement
In scope – handled inside Module 6:
| Concern | Examples |
|---|---|
| Provider authentication & credential storage | OAuth 2.0 flows, API-key vaults, token refresh |
| Request construction & serialisation | REST, GraphQL, SOAP envelope building |
| Response parsing & normalisation | Mapping provider-specific DTOs to internal canonical models |
| Retry, back-off & circuit-breaker logic | Exponential retry, dead-letter queues, half-open probes |
| Rate-limit management | Leaky-bucket tracking, queue throttling, cost estimation |
| Webhook ingestion & verification | HMAC signature checks, idempotent handler dispatch |
| Idempotency framework | Dedup window, idempotency-key generation, record storage |
| Provider health monitoring | Heartbeat checks, latency histograms, uptime SLAs |
Out of scope – remains in the originating module:
| Concern | Owning Module |
|---|---|
| Deciding when to sync a product listing | Module 3 (Catalog) |
| Deciding which items need restocking from a channel | Module 4 (Inventory) |
| Applying business rules to a payment result (e.g., partial-capture policy) | Module 1 (Sales) |
| Configuring which integrations are enabled per tenant | Module 5 (Setup) |
| Rendering integration status in the Admin Portal UI | Module 5 (Setup) / Front-end layer |
6.1.3 Cross-References to Source Modules
Cross-Reference: See Module 1, Section 1.6 for payment-processing business rules that Module 6 executes against the configured payment provider.
Cross-Reference: See Module 3, Section 3.9 for the product-sync lifecycle that triggers Module 6 outbound calls to Shopify, Amazon, and Google Merchant.
Cross-Reference: See Module 4, Section 4.5 for inventory-level sync events that Module 6 pushes to connected channels.
Cross-Reference: See Module 5, Section 5.4 for tenant-level integration configuration (credentials, enabled providers, sync schedules).
6.1.4 Integration Types
| Code | Provider | Direction | Primary Data |
|---|---|---|---|
SHOPIFY | Shopify (REST + GraphQL) | Bi-directional | Products, orders, inventory levels, fulfillments |
AMAZON | Amazon SP-API | Bi-directional | Listings, orders, FBA inventory, reports |
GOOGLE_MERCHANT | Google Merchant Center | Outbound + webhooks | Product feeds, local inventory, promotions |
PAYMENT_PROCESSOR | Stripe / Square / Adyen | Bi-directional | Charges, refunds, disputes, payouts |
EMAIL_PROVIDER | SendGrid / Postmark / SES | Outbound + webhooks | Transactional email, delivery events |
SHIPPING_CARRIER | EasyPost / ShipStation | Bi-directional | Rate quotes, labels, tracking events |
6.1.5 Module Interconnection Diagram
flowchart TD
M1[Module 1: Sales] -->|Payment flow| M6
M3[Module 3: Catalog] -->|Product sync| M6
M4[Module 4: Inventory] -->|Stock sync| M6
M5[Module 5: Setup] -->|Configuration| M6
M6[Module 6: Integrations]
M6 -->|Product listings| SHOP[Shopify]
M6 -->|Product listings| AMZ[Amazon SP-API]
M6 -->|Local inventory| GOOG[Google Merchant]
M6 -->|Card processing| PAY[Payment Processors]
M6 -->|Transactional email| EMAIL[Email Providers]
M6 -->|Rate/Label/Track| CARRIER[Shipping Carriers]
6.2 Integration Architecture
6.2.1 Provider Abstraction Layer
Every external provider – regardless of protocol – implements a common interface so that calling code never depends on provider-specific details.
Abstract Interface: IIntegrationProvider
public interface IIntegrationProvider
{
Task<ConnectionResult> Connect(ProviderCredentials credentials);
Task<DisconnectResult> Disconnect(string connectionId);
Task<SyncResult> Sync(SyncRequest request);
Task<ProviderStatus> GetStatus(string connectionId);
Task<ValidationResult> ValidateCredentials(ProviderCredentials credentials);
}
| Method | Purpose | Idempotent | Timeout |
|---|---|---|---|
Connect | Establish a new authenticated session or exchange OAuth tokens | No | 30 s |
Disconnect | Revoke tokens and mark the connection inactive | Yes | 15 s |
Sync | Execute a data-synchronisation operation (push or pull) | Yes (via idempotency key) | 120 s |
GetStatus | Return current health, last-sync timestamp, error counts | Yes | 10 s |
ValidateCredentials | Test credentials without persisting a connection | Yes | 15 s |
Provider Registry Pattern
The system maintains a runtime registry of all available provider implementations. At startup, each provider self-registers via dependency injection. Tenant configuration (Module 5) determines which providers are enabled for a given tenant at runtime.
# Example provider registry configuration
integration:
providers:
shopify:
enabled: true
implementation: ShopifyProvider
max_connections_per_tenant: 3
amazon:
enabled: true
implementation: AmazonSpApiProvider
max_connections_per_tenant: 2
google_merchant:
enabled: true
implementation: GoogleMerchantProvider
max_connections_per_tenant: 1
payment_processor:
enabled: true
implementation: StripeProvider # swappable
max_connections_per_tenant: 2
email_provider:
enabled: true
implementation: SendGridProvider # swappable
max_connections_per_tenant: 1
shipping_carrier:
enabled: true
implementation: EasyPostProvider # swappable
max_connections_per_tenant: 5
Data Model: integration_providers
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants.id |
provider_type | VARCHAR(50) | Yes | Enum: SHOPIFY, AMAZON, GOOGLE_MERCHANT, PAYMENT_PROCESSOR, EMAIL_PROVIDER, SHIPPING_CARRIER |
provider_name | VARCHAR(100) | Yes | Human-readable name (e.g., “Nexus Shopify Store”) |
status | VARCHAR(20) | Yes | ACTIVE, INACTIVE, ERROR, PENDING_AUTH |
credentials_vault_ref | VARCHAR(255) | Yes | Reference to encrypted credential store (never plaintext) |
config_json | JSONB | No | Provider-specific configuration overrides |
last_sync_at | TIMESTAMPTZ | No | Timestamp of most recent successful sync |
last_error_at | TIMESTAMPTZ | No | Timestamp of most recent error |
error_count | INTEGER | Yes | Rolling error count (reset on success) – default 0 |
circuit_state | VARCHAR(15) | Yes | CLOSED, OPEN, HALF_OPEN – default CLOSED |
created_at | TIMESTAMPTZ | Yes | Row creation timestamp |
updated_at | TIMESTAMPTZ | Yes | Last modification timestamp |
Cross-Reference: See Module 5, Section 5.4.3 for the admin UI screens that manage rows in this table.
6.2.2 Authentication Patterns
Each provider family uses a different authentication mechanism. The table below summarises the patterns; subsequent subsections detail the flows.
| Provider | Auth Method | Token Lifetime | Refresh Mechanism | Credential Storage |
|---|---|---|---|---|
| Shopify | OAuth 2.0 (offline access tokens) | Permanent (until revoked) | N/A – token does not expire | Encrypted vault, single token per store |
| Amazon SP-API | OAuth 2.0 via Login with Amazon (LWA) | 1 hour | refresh_token grant to LWA endpoint | Vault stores refresh_token; access token cached in memory |
| Google Merchant | OAuth 2.0 (Service Account) | 1 hour | Self-signed JWT assertion exchanged for access token | Vault stores service-account JSON key file |
| Payment Processor | API Key + Secret | Permanent (until rotated) | Manual key rotation via provider dashboard | Vault stores key pair; rotation tracked in audit log |
| Email Provider | API Key or SMTP credentials | Permanent (until rotated) | Manual key rotation | Vault stores API key or SMTP user/pass |
| Shipping Carrier | API Key | Permanent (until rotated) | Manual key rotation | Vault stores API key |
Token refresh is automatic. For providers with expiring tokens (Amazon, Google), the Integration Module schedules a background refresh 5 minutes before expiry. If the refresh fails, the circuit breaker transitions the provider to ERROR state and the retry mechanism takes over.
Cross-Reference: See Module 5, Section 5.4.5 for the credential-rotation workflow that triggers re-validation of stored credentials.
6.2.3 Retry & Backoff Strategy
All outbound calls from Module 6 pass through a shared retry pipeline. The strategy uses exponential backoff with jitter to avoid thundering-herd effects when a provider recovers from an outage.
Retry Parameters
| Parameter | Value | Notes |
|---|---|---|
| Max retries | 3 | After initial attempt |
| Base delay | 5 seconds | First retry wait |
| Multiplier | 3x | 5 s -> 15 s -> 45 s |
| Jitter | +/- 20 % | Randomised to spread load |
| Retryable status codes | 429, 500, 502, 503, 504 | Non-retryable: 400, 401, 403, 404, 409 |
| Dead-letter after | 3 failed retries | Message persisted for manual review |
Retry Sequence Diagram
sequenceDiagram
participant Caller as Business Module
participant RM as Retry Manager
participant Provider as External Provider
participant DLQ as Dead Letter Queue
Caller->>RM: Send request
RM->>Provider: Attempt 1
Provider-->>RM: 503 Service Unavailable
Note over RM: Wait 5s (+/- jitter)
RM->>Provider: Attempt 2
Provider-->>RM: 503 Service Unavailable
Note over RM: Wait 15s (+/- jitter)
RM->>Provider: Attempt 3
Provider-->>RM: 503 Service Unavailable
Note over RM: Wait 45s (+/- jitter)
RM->>Provider: Attempt 4 (final)
Provider-->>RM: 503 Service Unavailable
RM->>DLQ: Persist failed message
RM-->>Caller: SyncResult { Success: false, Reason: "DLQ" }
Dead Letter Queue Processing
Failed messages land in the integration_dead_letters table. An operations dashboard (Module 5 Admin Portal) surfaces these records. Operators can:
- Retry – re-enqueue the message for another attempt cycle.
- Skip – mark as resolved without retrying (e.g., stale data that has since been corrected).
- Escalate – flag for engineering investigation.
6.2.4 Circuit Breaker Pattern
The circuit breaker prevents a failing provider from consuming retry budget and delaying upstream callers. Each integration_providers row maintains its own circuit state.
State Machine
| Current State | Condition | Next State | Action |
|---|---|---|---|
CLOSED | Fewer than 5 failures in 60 s | CLOSED | Pass requests through normally |
CLOSED | 5 or more failures in 60 s | OPEN | Reject all requests immediately; log alert |
OPEN | Less than 30 s since transition | OPEN | Return cached error; do not attempt call |
OPEN | 30 s cooldown elapsed | HALF_OPEN | Allow a single probe request |
HALF_OPEN | Probe succeeds | CLOSED | Reset failure counter; resume normal traffic |
HALF_OPEN | Probe fails | OPEN | Restart 30 s cooldown; increment alert severity |
stateDiagram-v2
[*] --> CLOSED
CLOSED --> OPEN : 5 failures in 60s
OPEN --> HALF_OPEN : 30s cooldown elapsed
HALF_OPEN --> CLOSED : Probe succeeds
HALF_OPEN --> OPEN : Probe fails
CLOSED --> CLOSED : Request succeeds / failures < threshold
Circuit Breaker Configuration
circuit_breaker:
failure_threshold: 5
failure_window_seconds: 60
cooldown_seconds: 30
half_open_max_probes: 1
alert_on_open: true
alert_channel: "ops-integrations"
Cross-Reference: See Module 5, Section 5.18 for the alerting configuration that fires when a circuit opens.
6.2.5 Idempotency Framework
All mutating operations (product creates, inventory updates, payment captures) are wrapped in an idempotency layer. This guarantees exactly-once semantics even when retries or duplicate webhooks occur.
Dedup Window
| Parameter | Value |
|---|---|
| Window duration | 24 hours |
| Key algorithm | SHA-256 |
| Key input | provider + entity_type + entity_id + operation + timestamp_bucket |
| Timestamp bucket | Truncated to nearest hour |
| Cleanup schedule | Daily at 03:00 UTC – purge records older than 48 hours |
Idempotency Key Generation
idempotency_key = SHA-256(
provider = "SHOPIFY"
entity_type = "PRODUCT"
entity_id = "prod_abc123"
operation = "UPDATE"
timestamp_bucket = "2026-02-17T14:00:00Z" // truncated to hour
)
If a record with the same key already exists in the idempotency_records table, the system returns the cached response without re-executing the operation.
Data Model: idempotency_records
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants.id |
idempotency_key | VARCHAR(64) | Yes | SHA-256 hex digest (unique index) |
provider_type | VARCHAR(50) | Yes | Provider enum value |
entity_type | VARCHAR(50) | Yes | E.g., PRODUCT, ORDER, INVENTORY_LEVEL |
entity_id | VARCHAR(100) | Yes | Domain entity identifier |
operation | VARCHAR(30) | Yes | CREATE, UPDATE, DELETE, SYNC |
request_hash | VARCHAR(64) | Yes | SHA-256 of the serialised request body |
response_status | INTEGER | No | HTTP status code of cached response |
response_body | JSONB | No | Cached response payload |
executed_at | TIMESTAMPTZ | Yes | When the operation was executed |
expires_at | TIMESTAMPTZ | Yes | executed_at + 24 hours |
created_at | TIMESTAMPTZ | Yes | Row creation timestamp |
6.2.6 Rate Limit Management
Each provider enforces its own rate limits. Module 6 tracks consumption in real time and throttles outbound requests to stay within published quotas.
Rate Limits by Provider
| Provider | Limit Type | Rate | Strategy |
|---|---|---|---|
| Shopify REST | Leaky bucket | 40-request bucket, leaks at 2 req/s | Queue with inter-request delay; monitor X-Shopify-Shop-Api-Call-Limit header |
| Shopify GraphQL | Calculated query cost | 1,000-point bucket; restores at 50 pts/s (regular) or 500 pts/s (Plus) | Pre-calculate query cost; defer if bucket < required cost |
| Amazon SP-API | Token bucket | Per-endpoint (e.g., getOrders 0.0167 req/s burst 20) | Read x-amzn-RateLimit-Limit response header; adjust dynamically |
| Google Merchant | Daily quota | Products.insert: 2 calls/product/day; batch up to 10,000 entries | Batch mutations into single custombatch call; schedule during off-peak |
| Payment Processor (Stripe) | Sliding window | 100 read/s, 100 write/s (live mode) | Token bucket in-memory; back-off on 429 |
| Email Provider (SendGrid) | Sliding window | Varies by plan (100-10,000 emails/s) | In-memory counter; degrade gracefully |
| Shipping Carrier (EasyPost) | Per-second | 50 req/s | In-memory token bucket |
Rate Limit Tracking
The module maintains an in-memory sliding-window counter per (tenant_id, provider_type) pair. When the counter approaches the limit (80 % threshold), requests are queued rather than rejected. If the queue depth exceeds 500 items, new requests receive a 429 back-pressure signal to the calling module.
rate_limiter:
warning_threshold_pct: 80
max_queue_depth: 500
queue_drain_interval_ms: 100
metrics_export_interval_s: 10
6.2.7 Webhook Processing Pipeline
Inbound webhooks from external providers flow through a multi-stage pipeline that verifies authenticity, deduplicates, and dispatches to the appropriate handler.
Signature Verification by Provider
| Provider | Verification Method | Header / Field | Algorithm |
|---|---|---|---|
| Shopify | HMAC signature | X-Shopify-Hmac-Sha256 | HMAC-SHA256 of raw body with app secret |
| Amazon SP-API | SQS message validation | SignatureVersion, Signature | RSA-SHA1 of canonical message string using Amazon certificate |
| Google Merchant | Push endpoint SSL + JWT | Authorization: Bearer <token> | Verify JWT against Google public keys; validate audience claim |
| Stripe | HMAC signature | Stripe-Signature | HMAC-SHA256 with whsec_ signing secret; timestamp tolerance 300 s |
| SendGrid | Basic auth or OAuth | Authorization header | Compare against stored webhook signing key |
| EasyPost | HMAC signature | X-Hmac-Signature | HMAC-SHA256 of raw body with webhook secret |
Webhook Processing Flow
sequenceDiagram
participant EXT as External Provider
participant GW as API Gateway
participant SIG as Signature Verifier
participant DEDUP as Idempotency Check
participant HANDLER as Webhook Handler
participant DLQ as Dead Letter Queue
participant BUS as Domain Event Bus
EXT->>GW: POST /webhooks/{provider}
GW->>SIG: Verify signature
alt Signature invalid
SIG-->>GW: 401 Unauthorized
GW-->>EXT: 401
else Signature valid
SIG->>DEDUP: Check idempotency key
alt Duplicate webhook
DEDUP-->>GW: 200 OK (cached)
GW-->>EXT: 200
else New webhook
DEDUP->>HANDLER: Dispatch to typed handler
alt Handler succeeds
HANDLER->>BUS: Publish domain event
HANDLER-->>DEDUP: Cache result
DEDUP-->>GW: 200 OK
GW-->>EXT: 200
else Handler fails
HANDLER->>DLQ: Persist failed webhook
HANDLER-->>GW: 200 OK (accepted for retry)
GW-->>EXT: 200
end
end
end
Idempotent Handler Pattern
Every webhook handler follows this template:
- Extract idempotency key from the webhook payload (e.g., Shopify’s
X-Shopify-Webhook-Id, Amazon’snotificationId, Stripe’sevent.id). - Check
idempotency_records– if a record exists and has not expired, return the cached response immediately. - Execute business logic – map the webhook payload to a domain event and publish it to the internal event bus.
- Persist idempotency record – store the key and response so future duplicates are short-circuited.
- Return 200 – always return
200 OKto the provider promptly. Processing failures are handled asynchronously via the dead-letter queue. Returning non-200codes to providers like Shopify triggers exponential re-delivery, which compounds the problem.
Dead Letter Queue for Webhooks
Failed webhooks are stored in integration_dead_letters with full payload and error context. The operations dashboard allows:
| Action | Description |
|---|---|
| Replay | Re-inject the webhook into the processing pipeline |
| Discard | Mark as resolved; no further action |
| Investigate | Attach notes; assign to engineering team |
Cross-Reference: See Module 5, Section 5.18.4 for the admin UI that manages dead-letter webhook records.
6.3 Shopify Integration (Enhanced)
Scope: End-to-end synchronization of product catalog, inventory, orders, and customer data between the Nexus POS system and Shopify e-commerce. This section consolidates all Shopify integration requirements previously distributed across Module 3 (Section 3.7 – Catalog Sync), Module 4 (Section 4.14.3 – Inventory Sync), and Module 5 (Section 5.16.3 – Integration Configuration), and adds new material covering GraphQL API preferences, bulk operations, idempotency, rate limits, webhook topics, POS UI extensions, hardware compatibility, omnichannel requirements, and third-party POS compliance rules.
Cross-Reference: See Section 6.1 for the Integration Hub architecture. See Section 6.2 for integration credentials and security model. See Module 3, Section 3.6 for multi-channel management. See Module 4, Section 4.14 for online order fulfillment workflows.
Consolidation Note: This section replaces and supersedes the following legacy sections:
- Old Section 3.7 (Shopify Integration) – catalog sync modes, field ownership, conflict resolution
- Old Section 4.14.3 (Inventory Sync with Shopify) – inventory sync triggers and architecture
- Old Section 5.16.3 (Shopify Integration) – configuration fields, webhook endpoints
6.3.1 Sync Modes
Two tenant-configurable modes control how data flows between the POS and Shopify:
| Mode | Direction | Default | Use Case |
|---|---|---|---|
| POS-Master | POS -> Shopify (one-way) | Yes | All product data managed in POS. Changes in Shopify for POS-owned fields are overwritten on next sync. Recommended for retailers who manage their catalog exclusively through the POS. |
| Bidirectional with POS Priority | POS <-> Shopify (two-way) | No | Changes flow both directions. POS-owned fields: POS wins on conflict. Shopify-owned fields: Shopify wins. Configurable fields: POS wins with audit trail. Supports retailers whose external teams (SEO agencies, product photographers, copywriters) work directly in Shopify Admin. |
Rationale: Industry standard (Lightspeed, Retail Pro, SKU IQ) uses POS-master as default. Bidirectional supports retailers whose external teams work directly in Shopify Admin. ADR #24 documents this decision.
Important: Inventory sync is ALWAYS bidirectional regardless of catalog sync mode. Even in POS-Master mode, Shopify online sales decrement POS inventory and POS sales decrement Shopify inventory.
Sync Decision Flowchart
flowchart TD
A[Change Detected] --> B{Which System?}
B -->|POS Change| C{Field Ownership?}
B -->|Shopify Change| D{Field Ownership?}
C -->|POS-Owned| E[Push to Shopify via GraphQL Mutation]
C -->|Shopify-Owned| F[Ignore -- Not POS Managed]
C -->|Configurable| G{Sync Mode?}
D -->|POS-Owned| H{Sync Mode?}
D -->|Shopify-Owned| I[Pull to POS as Read-Only]
D -->|Configurable| J{Sync Mode?}
G -->|POS-Master| E
G -->|Bidirectional| E
H -->|POS-Master| K[Overwrite on Next Sync]
H -->|Bidirectional| L[Reject Shopify Change + Log Conflict]
J -->|POS-Master| K
J -->|Bidirectional| M{Conflict?}
M -->|Same Field Changed Both Sides| N[POS Wins + Log Audit Entry]
M -->|Different Fields Changed| O[Merge Both Changes]
E --> P[Log Sync Event to Integration Sync Log]
I --> P
K --> P
L --> P
N --> P
O --> P
6.3.2 Field-Level Ownership
Three ownership categories determine which system has authority over each product field. This model eliminates the majority of sync conflicts by design.
| Category | Fields | Direction | Rationale |
|---|---|---|---|
| POS-Owned | SKU, barcode, base_price, cost, compare_at_price, variants (options/values), vendor, weight, dimensions, tax_code, product_type, lifecycle_status, inventory_qty | POS -> Shopify | Core retail operations data managed exclusively in POS. These fields drive pricing, costing, and inventory accuracy. |
| Shopify-Owned | SEO title (meta_title), SEO description (meta_description), URL handle (slug), metafields, additional images (after primary), image alt text, online collections, sales channel publishing | Shopify -> POS (read-only) | Online-only fields with no POS equivalent. Managed by e-commerce team or SEO agencies directly in Shopify Admin. |
| Configurable | long_description, tags, primary_image | Default: POS -> Shopify. Bidirectional when enabled. | May require external editing when retailer uses agencies for product content. Tenant can toggle per field. |
Business Rules:
- POS-Owned fields changed in Shopify (in POS-Master mode) are silently overwritten on the next sync cycle.
- Shopify-Owned fields are stored in POS as read-only reference data for display purposes (e.g., showing SEO title in admin dashboard).
- Configurable field direction is set per tenant in the Integration Configuration (Section 6.2).
- Field ownership is enforced at the API layer – the sync engine checks ownership before applying any inbound change.
6.3.3 Conflict Resolution
Per-field ownership eliminates the majority of sync conflicts. The remaining edge cases – primarily configurable fields modified on both sides in bidirectional mode – are handled by the conflict resolution engine.
Conflict Audit Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
product_id | UUID | Yes | FK to products table – product where conflict occurred |
shopify_product_id | String(50) | Yes | Shopify product GID (e.g., gid://shopify/Product/123456) |
field_name | String(100) | Yes | The specific field in conflict (e.g., long_description, tags, primary_image) |
pos_value | Text | Yes | Value from POS at time of conflict |
shopify_value | Text | Yes | Value from Shopify at time of conflict |
resolved_value | Text | Yes | Final value written to both systems |
resolution_method | Enum | Yes | AUTO_POS_WINS, AUTO_SHOPIFY_WINS, AUTO_MERGE, MANUAL |
resolved_by | UUID | No | FK to users table – user who manually resolved (null for auto-resolution) |
resolved_at | DateTime | Yes | Timestamp of resolution |
created_at | DateTime | Yes | Conflict detection timestamp |
Conflict Resolution Sequence
sequenceDiagram
autonumber
participant SH as Shopify Webhook
participant GW as Webhook Gateway
participant API as POS Sync Engine
participant DB as POS Database
participant Q as Conflict Review Queue
SH->>GW: Product Updated Webhook (HMAC-SHA256 signed)
GW->>GW: Verify HMAC Signature
GW->>API: Forward Verified Payload
API->>API: Parse Changed Fields from Webhook Body
API->>DB: Fetch Current POS Product State
loop For Each Changed Field
API->>API: Determine Field Ownership Category
alt POS-Owned Field Changed in Shopify
API->>DB: Log Conflict Audit (resolution = AUTO_POS_WINS)
API->>SH: Push POS Value Back to Shopify (GraphQL Mutation)
else Shopify-Owned Field
API->>DB: Update POS Record (read-only reference copy)
else Configurable Field (Bidirectional Mode)
API->>DB: Check POS last_modified Timestamp
alt POS Also Changed Same Field (within sync window)
API->>DB: POS Wins -- Log Conflict Audit
API->>SH: Push POS Value to Shopify
else Only Shopify Changed
API->>DB: Accept Shopify Value into POS
end
end
end
opt Manual Review Flagged
API->>Q: Add to Admin Review Queue
API->>DB: Set conflict status = PENDING_REVIEW
end
Business Rules:
- Configurable fields in bidirectional mode: POS value wins automatically. The overridden Shopify value is preserved in the conflict audit table.
- Non-conflicting changes (different fields modified on each side) merge automatically with no conflict entry created.
- Optional manual review queue: admin dashboard shows pending conflicts flagged for human review. Staff with ADMIN or OWNER role can override auto-resolution.
- Conflict audit records are retained for 12 months for compliance and debugging, then archived.
- Conflicts exceeding 10 per product per day trigger an alert to the admin dashboard (possible integration loop).
6.3.4 Sync Constraints & Technical Notes
| Constraint | Limit | Handling Strategy |
|---|---|---|
| Max variants per product | 100 (Shopify hard limit) | Products exceeding 100 variants are flagged and excluded from Shopify sync with admin notification. Tracked in Sync Coverage report. |
| Max option dimensions | 3 (Shopify hard limit) | POS supports 3 dimensions, aligned with Shopify. Products with >3 dimensions cannot be synced. |
| API rate limit (REST) | 40 req bucket (regular) / 400 req bucket (Plus) | Request queue with leaky bucket tracking; batch operations throttled to stay within budget. |
| API rate limit (GraphQL) | 50 points/sec (regular) / 500 points/sec (Plus) | Query cost pre-calculation before execution; throttle queue when approaching limit. |
| Webhook reliability | HMAC-SHA256 signature verification required | Idempotent handlers with deduplication key. Dead letter queue for failed processing. |
| Product title max length | 255 characters (Shopify limit) | Truncate with ellipsis and log warning. |
| Tag max count | 250 tags per product (Shopify limit) | Excess tags dropped with admin notification. |
Image Sync:
- POS sends the primary image on first product publish only.
- All subsequent image management is performed in Shopify.
- Additional images added in Shopify are pulled to POS as read-only references for display in the POS UI.
- Image URLs from Shopify CDN are stored – images are NOT downloaded to POS storage.
Inventory Sync:
- Inventory sync is ALWAYS bidirectional regardless of catalog sync mode.
- Sales on Shopify decrement POS inventory. POS sales decrement Shopify inventory.
- Uses Shopify Inventory API (
inventorySetQuantitiesGraphQL mutation) with location-level granularity. - Each POS physical location maps 1:1 to a Shopify location.
Sync Timing:
- Real-time for individual changes (webhook-driven + API push).
- Batch processing for bulk operations (imports, stock takes) queued and processed during off-peak hours.
- Periodic reconciliation job runs every 15 minutes to detect and correct drift.
- Daily full reconciliation at 02:00 local time as a safety net.
Variant Limit Handling:
- Products exceeding 100 variants or 3 option dimensions are automatically excluded from Shopify sync.
- Admin receives notification: “Product [SKU] excluded from Shopify sync: exceeds variant limits.”
- Excluded products are tracked in the Sync Coverage report with exclusion reason.
6.3.5 Reports: Shopify Integration
| Report | Purpose | Key Data Fields |
|---|---|---|
| Sync Status Dashboard | Monitor overall sync health in real-time | Products synced, products pending, products failed, last sync time, error count (24h), average sync latency (ms) |
| Conflict Log | Review and resolve sync conflicts | Product SKU, field name, POS value, Shopify value, resolution method, timestamp, resolved by |
| Sync Coverage | Identify products NOT synced to Shopify | Product SKU, product name, exclusion reason (exceeds limits / manually excluded / new / sync error), action needed |
| Sync Activity Log | Detailed sync event history for troubleshooting | Product SKU, direction (POS->Shopify / Shopify->POS), fields synced, timestamp, duration (ms), result (success/failed/partial) |
Report Access: All reports available in Admin Portal under Integrations > Shopify. Exportable to CSV. Sync Status Dashboard refreshes every 60 seconds.
6.3.6 GraphQL API Preference & Query Cost Model
Shopify recommends GraphQL over REST for all new development. The Nexus POS integration MUST use the Shopify GraphQL Admin API as the primary interface.
Rationale:
- GraphQL returns only requested fields, reducing payload size and bandwidth.
- Single request can fetch related resources (product + variants + images) without multiple round trips.
- GraphQL supports bulk operations not available in REST.
- Shopify is actively deprecating REST endpoints in favor of GraphQL.
Query Cost Calculation:
Shopify calculates query cost using the formula:
cost = requested_fields * requested_objects
Each query returns an extensions.cost object with requestedQueryCost, actualQueryCost, and throttleStatus (including currentlyAvailable points).
Query Cost Budget:
| Plan | Points per Second | Max Single Query Cost | Restore Rate |
|---|---|---|---|
| Regular (Development/Basic/Shopify/Advanced) | 50 | 1,000 | 50 points/sec |
| Shopify Plus | 500 | 10,000 | 500 points/sec |
Common Operation Costs:
| Operation | Estimated Cost (points) | Notes |
|---|---|---|
| Fetch single product with variants | 12-15 | Depends on variant count |
| Fetch 50 products (list) | 50-80 | Paginated with cursor |
| Update single product | 10 | productUpdate mutation |
| Update inventory at location | 10 | inventorySetQuantities mutation |
| Fetch 250 inventory levels | 30-50 | inventoryItems query with locations |
| Create product with 5 variants | 15-20 | productCreate mutation |
| Bulk product query (JSONL) | 10 (submit) | Actual processing is async |
Implementation Rules:
- All Shopify API calls MUST use GraphQL Admin API (version
2025-04or later). - REST API is permitted only for endpoints not yet available in GraphQL (e.g., certain carrier service endpoints).
- Every query MUST request the
extensions.costfield to monitor budget consumption. - Sync engine MUST track
currentlyAvailablepoints and pause requests when below 20% of maximum budget. - Query complexity MUST be pre-estimated before execution; queries exceeding 80% of max single query cost must be split.
6.3.7 Idempotency Directive
Starting with API version 2026-04, Shopify requires the @idempotent directive on all mutations to prevent duplicate operations during retries and network failures.
Implementation:
All Shopify mutations issued by the POS sync engine MUST include an idempotencyKey parameter:
idempotency:
key_generation:
algorithm: SHA-256
input_template: "{tenant_id}:{mutation_name}:{entity_id}:{timestamp_bucket}"
timestamp_bucket: 5_minutes # Floor timestamp to 5-minute windows
retry_behavior:
same_key_returns: cached_response
cache_ttl: 60_minutes # Shopify caches idempotent responses for 60 min
example:
tenant_id: "t_abc123"
mutation: "productUpdate"
entity_id: "gid://shopify/Product/789"
timestamp_bucket: "2026-02-17T14:30" # Floored to 5-min boundary
key: "SHA256(t_abc123:productUpdate:gid://shopify/Product/789:2026-02-17T14:30)"
Business Rules:
- Every mutation wrapper function MUST generate and attach an idempotency key before submission.
- If a mutation fails with a network timeout, the sync engine retries with the SAME idempotency key. Shopify returns the cached result if the original mutation succeeded.
- Idempotency keys are logged in the Integration Sync Log for audit and debugging.
- The 5-minute timestamp bucket prevents accidental deduplication of legitimate rapid updates (e.g., price change followed by description change within seconds).
6.3.8 Bulk Operations API
For large data syncs (initial catalog onboarding, full inventory reconciliation, daily product exports), the POS system uses Shopify’s Bulk Operations API instead of individual GraphQL queries.
Bulk Operation Lifecycle:
sequenceDiagram
autonumber
participant POS as POS Sync Engine
participant GQL as Shopify GraphQL API
participant S3 as Shopify Storage (JSONL)
POS->>GQL: Submit bulkOperationRunQuery / bulkOperationRunMutation
GQL-->>POS: Return operation ID + status: CREATED
loop Poll for Completion (every 10 seconds)
POS->>GQL: query { currentBulkOperation { status url } }
GQL-->>POS: status: RUNNING / COMPLETED / FAILED
end
alt Operation Completed
GQL-->>POS: status: COMPLETED, url: "https://storage.shopify.com/result.jsonl"
POS->>S3: Download JSONL Result File
S3-->>POS: Stream JSONL Data
POS->>POS: Parse JSONL, Update Local Database
else Operation Failed
GQL-->>POS: status: FAILED, error details
POS->>POS: Log Error, Queue for Retry
end
Concurrency Limits:
| Operation Type | Max Concurrent | Scope |
|---|---|---|
| QUERY (bulk export) | 5 | Per app per shop |
| MUTATION (bulk import) | 5 | Per app per shop |
| Combined | 10 | 5 queries + 5 mutations simultaneously |
Use Cases:
| Use Case | Operation Type | Frequency | Estimated Duration |
|---|---|---|---|
| Initial catalog sync (new tenant onboarding) | QUERY (export all products from Shopify) | Once | 2-15 min depending on catalog size |
| Full inventory reconciliation | QUERY + MUTATION | Daily at 02:00 | 5-30 min |
| Daily product export to Shopify | MUTATION (staged uploads) | Daily at 03:00 | 5-20 min |
| Price book sync | MUTATION | On demand | 1-10 min |
| Variant bulk update | MUTATION | On demand | 1-10 min |
Implementation Rules:
- Bulk operations MUST be used for any sync involving more than 50 products or 100 inventory levels.
- Individual GraphQL queries are used for real-time, single-entity syncs (e.g., sale completed, product updated).
- JSONL results MUST be streamed (not loaded fully into memory) to handle large catalogs.
- Failed bulk operations are retried up to 3 times with 5-minute intervals before alerting admin.
- Bulk operation results are stored locally for 7 days for debugging.
6.3.9 POS UI Extensions
Shopify POS provides extension points for embedded apps. The Nexus POS integration leverages these APIs for enhanced in-store experiences.
Supported POS Extension APIs:
| API | Minimum Version | Purpose | Nexus Usage |
|---|---|---|---|
| Camera API | POS v10.0+ | Access device camera for barcode scanning | Scan product barcodes within embedded Nexus app view on Shopify POS hardware |
| Translation API | POS v10.19+ | Multi-language support for POS UI | Support bilingual staff (English/Spanish) at retail locations |
| Session Token API | POS v9.0+ | Authenticate embedded app sessions | Secure communication between Nexus embedded app and POS backend without separate login |
| Cart API | POS v10.0+ | Read and modify the active POS cart | Inject custom line items, apply Nexus-managed discounts |
| Customer API | POS v10.0+ | Access customer data in POS context | Display unified customer profile with cross-channel purchase history |
Session Token Authentication:
session_token:
flow: "Shopify POS -> Embedded App"
mechanism: JWT
issued_by: Shopify
validated_by: POS Backend (verify with Shopify public key)
claims:
- iss: "https://{shop}.myshopify.com/admin"
- dest: "https://{shop}.myshopify.com"
- sub: "{staff_member_id}"
- exp: "{expiry_timestamp}"
token_refresh: automatic (before expiry)
no_separate_login: true # Staff authenticates via Shopify POS PIN
Business Rules:
- The Nexus POS app MUST function as a Shopify POS UI extension when deployed alongside Shopify POS hardware.
- Session tokens replace API key authentication for all POS-embedded interactions.
- Camera API usage requires explicit permission grant during app installation.
- Translation API strings are managed in the Nexus localization system and pushed to the extension.
6.3.10 Rate Limits 2026
Shopify enforces rate limits across all API surfaces. The POS sync engine MUST respect these limits to avoid request rejection (HTTP 429) and potential app throttling.
Rate Limit Summary:
| API Type | Regular Store | Shopify Plus | Burst Capacity | Leak Rate / Restore |
|---|---|---|---|---|
| REST Admin API | 40-request bucket | 400-request bucket | Full bucket available immediately | 2 req/sec (regular), 20 req/sec (Plus) |
| GraphQL Admin API | 50 points/sec | 500 points/sec | N/A (cost-based) | Cost-based, restored continuously |
| Bulk Operations | 5 concurrent queries + 5 concurrent mutations | Same | N/A | Poll-based completion |
| Storefront API | 100 req/sec per app per store | Same | N/A | Fixed rate |
| Webhook Delivery | No outbound limit from Shopify | Same | N/A | Must respond within 5 seconds |
Throttle Handling Strategy:
rate_limit_strategy:
monitoring:
track_remaining_points: true # From X-Shopify-Shop-Api-Call-Limit header (REST) or extensions.cost (GraphQL)
alert_threshold: 20_percent # Alert when remaining capacity drops below 20%
backoff:
initial_delay_ms: 500
max_delay_ms: 30000
multiplier: 2.0
jitter: true # Add random jitter to prevent thundering herd
queue:
max_queue_size: 1000
priority_levels:
- CRITICAL: inventory_updates, order_syncs # Process first
- NORMAL: product_updates, customer_syncs
- LOW: bulk_exports, reconciliation
overflow_action: reject_with_retry_after
Business Rules:
- All API calls MUST check remaining rate limit budget before execution.
- When rate limited (HTTP 429), the sync engine MUST use the
Retry-Afterheader value for backoff timing. - Critical operations (inventory updates after sales) receive priority queue placement over non-critical operations (product description syncs).
- Rate limit consumption is logged per minute for capacity planning and plan upgrade recommendations.
6.3.11 Webhook Topics Catalog
The POS system registers for the following Shopify webhook topics. All webhooks are verified using HMAC-SHA256 before processing.
Registered Webhook Topics:
| Topic | Direction | Trigger | POS Action |
|---|---|---|---|
orders/create | Shopify -> POS | Customer places online order | Create fulfillment task at assigned store. Reserve inventory. Notify store staff. |
orders/updated | Shopify -> POS | Order modified (address change, note added) | Update local order record. Refresh fulfillment queue. |
orders/cancelled | Shopify -> POS | Online order cancelled | Release inventory reservation. Remove from fulfillment queue. Log cancellation. |
orders/fulfilled | Shopify -> POS | Order marked fulfilled (may be by third-party) | Update local order status. Confirm inventory decrement. |
orders/partially_fulfilled | Shopify -> POS | Partial shipment completed | Update line-item fulfillment status. Adjust remaining reservation. |
products/create | Shopify -> POS | Product created in Shopify Admin | Import as read-only reference (bidirectional mode). Ignore in POS-Master mode. |
products/update | Shopify -> POS | Product fields modified in Shopify | Apply field-ownership conflict resolution (Section 6.3.3). |
products/delete | Shopify -> POS | Product deleted from Shopify | Flag local product as SHOPIFY_DELETED. Do NOT delete from POS. Admin notification. |
inventory_levels/update | Shopify -> POS | Inventory adjusted in Shopify Admin or by another app | Sync adjusted quantity to POS at corresponding location. Log as EXTERNAL_ADJUSTMENT. |
inventory_levels/connect | Shopify -> POS | Product stocked at a new Shopify location | Create location-level inventory record in POS if location is mapped. |
inventory_levels/disconnect | Shopify -> POS | Product removed from a Shopify location | Set POS inventory to zero at that location. Flag for admin review. |
customers/create | Shopify -> POS | New customer registers online | Create or link customer profile in POS. Merge if email/phone matches existing. |
customers/update | Shopify -> POS | Customer profile updated | Sync updated fields to POS customer record (email, phone, address). |
refunds/create | Shopify -> POS | Online order refunded | Process refund in POS. Increment inventory if items restocked. Update order status. |
app/uninstalled | Shopify -> POS | Merchant uninstalls Nexus app | Disable all sync operations. Preserve local data. Alert admin. Set integration status to DISCONNECTED. |
shop/update | Shopify -> POS | Shop settings changed (currency, timezone, name) | Update cached shop metadata. Validate currency alignment. |
bulk_operations/finish | Shopify -> POS | A bulk operation completes | Download and process JSONL result file. Update sync status. |
Webhook Security & Reliability:
| Parameter | Value |
|---|---|
| Verification method | HMAC-SHA256 using app secret as key |
| Header | X-Shopify-Hmac-Sha256 (Base64-encoded) |
| Mandatory response time | < 5 seconds (return HTTP 200/201/202) |
| Shopify retry policy | 19 retries over 48 hours with exponential backoff |
| Failure threshold | After 19 failed deliveries, webhook is automatically removed by Shopify |
| Deduplication | X-Shopify-Webhook-Id header used as idempotency key for handler deduplication |
| Processing model | Acknowledge immediately (HTTP 200), process asynchronously via background job queue |
Business Rules:
- Webhook handlers MUST return HTTP 200 within 5 seconds. All processing happens asynchronously after acknowledgment.
- Every incoming webhook is deduplicated using
X-Shopify-Webhook-Idbefore queuing for processing. - Failed webhook processing (after acknowledgment) is retried 3 times internally before moving to dead letter queue.
- Webhook registrations are verified daily; any missing registrations are automatically re-created.
- All webhook payloads are logged (with PII redaction) for 30 days for debugging.
6.3.12 Third-Party POS Integration Rules
The Nexus POS is a non-native, custom POS system connecting to Shopify as a third-party application. This imposes specific compliance requirements and architectural constraints that differ from Shopify’s own native POS product.
Integration Architecture:
integration_type: third_party_pos
authentication: OAuth 2.0 (mandatory -- no API key bypass)
app_listing: Shopify App Store (or custom/unlisted app for private deployment)
data_authority: POS is source of truth for products and inventory
checkout_model: Must NOT bypass standard Shopify checkout for online orders
Compliance Requirements:
| Requirement | Description | Implementation |
|---|---|---|
| OAuth Authentication | All API access MUST use OAuth 2.0 access tokens obtained through Shopify’s authorization flow. Direct API key authentication is not permitted for production apps. | Implement OAuth install flow with PKCE. Store encrypted access tokens in Integration Credentials table (Section 6.2). |
| App Security | App MUST follow Shopify’s mandatory security requirements including HTTPS, secure credential storage, and regular security audits. | TLS 1.2+ for all API calls. AES-256 encryption for stored tokens. Annual security review. |
| Checkout Integrity | Third-party POS MUST NOT bypass Shopify’s standard checkout process for online orders. In-store transactions process through the POS payment system. | Online orders flow through Shopify checkout. POS handles in-store payments via its own payment integration (Module 1, Section 1.18). |
| Data Overwrite Awareness | Data pushed from POS potentially overwrites Shopify data. The field ownership model (Section 6.3.2) controls which fields POS is authorized to modify. | Sync engine enforces field ownership before every write operation. Shopify-owned fields are never overwritten by POS. |
| Real-Time Integration | Must support real-time API integration to avoid data delays that cause overselling or price mismatches. | Webhook-driven architecture with < 5 second latency target. Scheduled reconciliation as safety net. |
| Partner Program Compliance | Must comply with Shopify Partner Program requirements including app review, privacy policy, and terms of service. | Maintain active Shopify Partner account. Submit app for review before merchant deployment. |
| Data Privacy | Must handle customer data according to Shopify’s data protection requirements. Implement data deletion webhooks. | Process customers/data_request, customers/redact, and shop/redact mandatory compliance webhooks. |
| API Versioning | Must use a supported API version (within 12 months of release). Deprecated versions result in app rejection. | Track Shopify API version calendar. Update to latest stable version within 6 months of release. Test against release candidate versions. |
Compliance Checklist:
| # | Item | Status | Verification Method |
|---|---|---|---|
| 1 | OAuth 2.0 implementation with PKCE | Required | Shopify app review |
| 2 | HTTPS on all endpoints (TLS 1.2+) | Required | SSL certificate check |
| 3 | HMAC webhook verification on all webhook handlers | Required | Code review |
| 4 | Mandatory compliance webhooks implemented (customers/data_request, customers/redact, shop/redact) | Required | Shopify app review |
| 5 | Access token encryption at rest (AES-256) | Required | Security audit |
| 6 | No Shopify checkout bypass for online orders | Required | Functional test |
| 7 | Field ownership enforcement (no unauthorized overwrites) | Required | Integration test suite |
| 8 | API version within supported window | Required | Automated version check |
| 9 | Privacy policy URL configured in app settings | Required | Shopify Partner Dashboard |
| 10 | App terms of service URL configured | Required | Shopify Partner Dashboard |
| 11 | Rate limit compliance (no brute-force retry loops) | Required | Log analysis |
| 12 | Idempotency keys on all mutations | Required (2026-04+) | Code review |
6.3.13 Shopify Sync Rules & Best Practices
This section documents the operational rules and best practices that govern the day-to-day Shopify integration. These rules reflect both Shopify platform requirements and Nexus POS architectural decisions.
Single Source of Truth
- POS is the inventory master (consistent with ADR #24: POS-Master default sync mode).
- Product data (titles, descriptions, variants, pricing) syncs FROM POS TO Shopify in the default mode.
- Inventory quantities ALWAYS sync bidirectionally regardless of catalog sync mode. This is non-negotiable – Shopify online sales must decrement POS inventory in real time.
Location Configuration
Every POS physical location maps 1:1 to a Shopify location. Online sales decrement inventory at the fulfillment location, NOT a global pool.
Location Mapping Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
pos_location_id | UUID | Yes | FK to locations table (POS location) |
shopify_location_id | String(50) | Yes | Shopify location GID (e.g., gid://shopify/Location/12345) |
shopify_location_name | String(100) | Yes | Shopify location display name (cached) |
is_fulfillment_location | Boolean | Yes | Whether this location fulfills online orders (default: true) |
is_active | Boolean | Yes | Whether sync is active for this location (default: true) |
last_synced_at | DateTime | No | Last successful inventory sync for this location |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Every active POS location MUST have a corresponding Shopify location before inventory sync is enabled.
- Location mapping is configured during tenant onboarding (Step 12 of onboarding wizard, Module 5, Section 5.20).
- If a POS location has no Shopify mapping, inventory changes at that location are NOT synced to Shopify. A warning is displayed in the Integration Health Dashboard.
- “Track Inventory” MUST be enabled for all synced products in Shopify. The sync engine verifies this during reconciliation and enables tracking automatically if missing.
Real-Time Sync Requirements
| Sync Event | Latency Target | Method |
|---|---|---|
| In-store sale completed -> Shopify inventory update | < 5 seconds | Direct GraphQL mutation (fire-and-forget with retry) |
| Shopify online order -> POS fulfillment queue | < 10 seconds | Webhook receipt + async processing |
| Product update in POS -> Shopify product update | < 30 seconds | Queued GraphQL mutation (batched for rate limit efficiency) |
| Customer profile change -> cross-system sync | < 60 seconds | Queued mutation (lower priority) |
| Full inventory reconciliation | Every 15 minutes | Scheduled comparison job |
| Full catalog reconciliation | Daily at 02:00 | Bulk operations API |
Omnichannel Requirements
Customer Profile Unification:
- In-store purchases are linked to the same Shopify customer profile using email or phone match.
- Customer merge logic (Module 2, Section 2.2) handles deduplication when an online customer first visits a physical store.
- Staff can view a customer’s complete purchase history (online + in-store) from the POS terminal.
Cross-Channel Returns:
- Items bought online can be returned in-store. The POS retrieves the Shopify order and processes the return locally.
- Items bought in-store can be returned via Shopify online return flow (if enabled by the retailer).
- Return policy engine (Module 1, Section 1.3) validates return eligibility regardless of purchase channel.
BOPIS (Buy Online, Pick Up In-Store):
sequenceDiagram
autonumber
participant C as Customer
participant SF as Shopify Storefront
participant WH as Shopify Webhook
participant POS as POS Backend
participant ST as Store Staff
participant N as Notification Service
C->>SF: Place Order (select "Local Pickup" at checkout)
SF->>SF: Create Order (fulfillment_status: unfulfilled, delivery_method: local_pickup)
SF->>WH: Emit orders/create webhook
WH->>POS: Deliver orders/create payload (HMAC verified)
POS->>POS: Identify assigned pickup location
POS->>POS: Reserve inventory at pickup location
POS->>ST: Display order on Fulfillment Queue (tagged: PICKUP)
POS->>N: Send "Order Ready for Prep" email to customer
ST->>POS: Pick items, mark as READY_FOR_PICKUP
POS->>N: Send "Your Order is Ready for Pickup" email to customer
POS->>SF: Update Shopify order (fulfillment: ready_for_pickup)
C->>ST: Arrive at store, present order confirmation
ST->>POS: Scan order barcode or search by order number
POS->>POS: Verify customer identity
ST->>POS: Mark as PICKED_UP (complete fulfillment)
POS->>SF: Update Shopify order (fulfillment_status: fulfilled)
POS->>POS: Decrement inventory (SALE movement type)
POS->>N: Send "Pickup Confirmed" email to customer
BOPIS Business Rules:
- BOPIS orders appear in the store’s fulfillment queue with a
PICKUPtag for visual distinction. - Pickup window is configurable per location (default: 48 hours). After expiry, staff is notified and order may be cancelled.
- Inventory is reserved (not decremented) until customer picks up. If pickup expires and order is cancelled, reservation is released.
- Customer receives three notifications: order confirmed, ready for pickup, pickup completed.
Staff & Security
| Requirement | Implementation |
|---|---|
| Staff PINs | 4-6 digit PINs for POS terminal access and sale attribution (Module 5, Section 5.5) |
| OAuth authentication | All Shopify API access uses OAuth tokens – no API key shortcuts in production |
| Cycle counts | Weekly recommended, monthly minimum. Keeps inventory accurate even with real-time auto-sync. Discrepancies logged and investigated. |
| Audit trail | Every sync operation, conflict resolution, and manual override is logged with user ID and timestamp |
| Token rotation | Access tokens are monitored for expiry. Re-authorization flow triggered 30 days before token expiration. |
6.3.14 Inventory Sync Triggers
Inventory quantities sync bidirectionally between the POS system and Shopify, regardless of catalog sync mode. This ensures online customers see accurate availability at all times.
Cross-Reference: See Module 4, Section 4.12 for the POS inventory reservation model (reserve on cart add, commit on sale complete). See Module 4, Section 4.14 for online order fulfillment and pick-pack-ship workflows.
POS -> Shopify Sync Triggers:
| POS Event | Shopify Update | Mutation Used |
|---|---|---|
| Sale completed | Decrement inventory at sale location | inventorySetQuantities |
| Return processed | Increment inventory at return location | inventorySetQuantities |
| Purchase order received | Increment inventory at receiving location | inventorySetQuantities |
| Inventory adjustment posted | Adjust inventory at location (increment or decrement) | inventorySetQuantities |
| Transfer completed | Decrement source location, increment destination location | inventorySetQuantities (two calls) |
| Stock count finalized | Set inventory to physical count result at location | inventorySetQuantities |
| Reservation expired | Release reserved quantity (no Shopify update – reservation is POS-only) | None |
Shopify -> POS Sync Triggers:
| Shopify Event | POS Update | Webhook Topic |
|---|---|---|
| Online order placed | Decrement available qty at assigned fulfillment store | orders/create |
| Online order cancelled | Release reservation; increment available qty | orders/cancelled |
| Online return processed | Increment available qty at return location | refunds/create |
| Manual Shopify adjustment | Sync to POS with EXTERNAL_ADJUSTMENT movement type | inventory_levels/update |
| Product stocked at new location | Create location-level inventory record | inventory_levels/connect |
| Product removed from location | Zero out inventory at that location | inventory_levels/disconnect |
Sync Architecture Parameters:
| Parameter | Value |
|---|---|
| Sync method | Webhook-driven (near real-time) + scheduled reconciliation |
| Webhook latency target | < 5 seconds from event to POS database update |
| Reconciliation frequency | Every 15 minutes (incremental) + daily at 02:00 (full) |
| Conflict resolution | POS is source of truth; POS value wins on discrepancy |
| Retry on failure | 3 retries with exponential backoff (5s, 15s, 45s) |
| Dead letter queue | Failed syncs after 3 retries queued for manual review |
| Reconciliation tolerance | Differences of +/- 0 units trigger correction (zero tolerance) |
Business Rule: Inventory sync does NOT depend on catalog sync mode. Even if a tenant uses POS-Master catalog sync (where POS owns product data), inventory quantities always flow bidirectionally. The POS system is the authoritative source for inventory levels – if a discrepancy is detected during reconciliation, the POS value overwrites the Shopify value.
6.3.15 Shopify Hardware Compatibility
When the Nexus POS is deployed alongside or integrated with Shopify POS hardware, the following devices are compatible. This section covers hardware that can be shared between Shopify POS and the Nexus POS application.
Card Readers:
| Device | Connection | Shopify POS | Nexus POS | Notes |
|---|---|---|---|---|
| Shopify Tap & Chip Reader | Bluetooth | Yes | No (uses own payment integration) | Shopify-exclusive; cannot process payments for third-party POS |
| Shopify POS Terminal (Chipper 2X BT) | Bluetooth | Yes | No | Shopify-exclusive hardware |
| WisePOS E (Stripe Terminal) | Internet / USB | No | Yes | Recommended for Nexus POS payment processing |
| BBPOS Chipper 2X BT | Bluetooth | Yes | Via Stripe SDK | Dual-compatible with configuration |
Note: Shopify-branded card readers process payments exclusively through Shopify Payments. The Nexus POS uses its own payment integration (Module 1, Section 1.18) with Stripe Terminal or equivalent semi-integrated devices. Card readers are NOT shared between the two POS systems.
Receipt Printers:
| Device | Connection | Protocol | Shopify POS | Nexus POS | Notes |
|---|---|---|---|---|---|
| Star Micronics TSP143IV | USB / LAN / Bluetooth | StarPRNT | Yes | Yes | Recommended. Shared between both POS systems. |
| Star Micronics mPOP | Bluetooth | StarPRNT | Yes | Yes | Combined printer + cash drawer. Compact form factor. |
| Star Micronics SM-L200 | Bluetooth | StarPRNT | Yes | Yes | Mobile receipt printer for floor sales. |
| Epson TM-m30III | USB / LAN / Bluetooth | ESC/POS | Yes | Yes | Alternative to Star Micronics. Industry standard protocol. |
| Epson TM-T88VII | USB / LAN | ESC/POS | No (not Shopify certified) | Yes | High-speed thermal printer. Nexus POS only. |
Barcode Scanners:
| Device | Connection | Shopify POS | Nexus POS | Notes |
|---|---|---|---|---|
| Socket Mobile S700 | Bluetooth | Yes | Yes | 1D barcode scanner. Compact, retail-grade. |
| Socket Mobile S740 | Bluetooth | Yes | Yes | 2D barcode scanner (supports QR codes). |
| Shopify Retail Scanner | Bluetooth | Yes | No | Shopify-exclusive accessory. |
| Any HID-compliant USB scanner | USB | Yes | Yes | Generic USB scanners work with both systems via keyboard wedge. |
| Zebra DS2208 | USB | No | Yes | High-performance 2D imager. Nexus POS recommended. |
Cash Drawers:
| Device | Connection | Shopify POS | Nexus POS | Notes |
|---|---|---|---|---|
| Star Micronics Cash Drawer (via mPOP) | Integrated | Yes | Yes | Opens via mPOP printer signal. |
| APG Vasario Cash Drawer | RJ-12 (via printer) | Yes | Yes | Standard cash drawer. Triggered by receipt printer kick signal. |
| Any RJ-12 compatible drawer | RJ-12 (via printer) | Yes | Yes | Opened by ESC/POS or StarPRNT printer kick command. |
Device Compatibility Matrix (Summary):
| Peripheral Type | Shared Between Shopify POS & Nexus POS | Nexus POS Only | Shopify POS Only |
|---|---|---|---|
| Card Readers | None | WisePOS E, Stripe Terminal devices | Shopify Tap & Chip, Chipper 2X BT |
| Receipt Printers | Star TSP143IV, Star mPOP, Star SM-L200, Epson TM-m30III | Epson TM-T88VII | None |
| Barcode Scanners | Socket Mobile S700/S740, USB HID scanners | Zebra DS2208 | Shopify Retail Scanner |
| Cash Drawers | All RJ-12 compatible (via shared printer) | None | None |
Business Rules:
- Receipt printers and cash drawers can be shared between Shopify POS and Nexus POS because they connect via standard protocols (ESC/POS, StarPRNT).
- Card readers are NOT shared – each POS system uses its own payment processing hardware.
- Barcode scanners using USB HID (keyboard wedge) mode work with any POS system that accepts keyboard input.
- Hardware compatibility is validated during tenant onboarding. Incompatible devices are flagged with recommended alternatives.
- The Nexus POS Register Management screen (Module 5, Section 5.7) tracks which peripherals are paired to each register.
6.4 Amazon SP-API Integration
Scope: Integrating the POS system with Amazon’s Selling Partner API (SP-API) to enable multi-channel retail operations across Amazon marketplaces. This section covers authentication, catalog synchronization, listing management, order fulfillment (FBA and FBM), inventory tracking, push notifications, rate limiting, compliance requirements, and reporting.
Cross-Reference: See Module 3, Section 3.7 for Shopify integration patterns (field ownership, conflict resolution, sync modes). See Module 4, Section 4.14 for the pick-pack-ship fulfillment workflow reused by Amazon FBM orders. See Module 5, Section 5.16 for the Integration Hub registry and health dashboard where Amazon is registered as an integration provider.
6.4.1 Authentication & Authorization
Amazon SP-API uses OAuth 2.0 via Login with Amazon (LWA) for all API access. The POS system acts as a registered SP-API application and stores per-tenant seller credentials.
OAuth 2.0 Token Flow:
- Access tokens are valid for 1 hour and must be refreshed using the
refresh_tokengrant type before expiry. - The POS backend handles all token lifecycle management transparently – store staff and admin users never interact with OAuth directly.
- Regional endpoints determine which Amazon marketplace cluster the API calls target.
Regional Endpoints:
| Region | Endpoint | Marketplaces |
|---|---|---|
| North America (NA) | sellingpartnerapi-na.amazon.com | US (ATVPDKIKX0DER), CA (A2EUQ1WTGCTBG2), MX (A1AM78C64UM0Y8) |
| Europe (EU) | sellingpartnerapi-eu.amazon.com | UK (A1F83G8C2ARO7P), DE (A1PA6795UKMFR9), FR, IT, ES |
| Far East (FE) | sellingpartnerapi-fe.amazon.com | JP (A1VC38T7YXB528), AU (A39IBJ37TRP1C6) |
Amazon Credential Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key, system-generated |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
integration_id | UUID | Yes | FK to integrations table (Module 5, Section 5.16) |
selling_partner_id | String(50) | Yes | Amazon Seller Central account identifier |
marketplace_id | String(20) | Yes | Amazon marketplace identifier (e.g., ATVPDKIKX0DER for US) |
client_id | String(100) | Yes | LWA application client ID |
client_secret_encrypted | String(500) | Yes | AES-256 encrypted LWA client secret. Never returned in API responses. |
refresh_token_encrypted | String(500) | Yes | AES-256 encrypted OAuth refresh token. Long-lived credential. |
access_token_encrypted | String(1000) | No | AES-256 encrypted current access token. Null when expired. |
token_expiry | DateTime | No | Expiration timestamp of the current access token |
region | Enum | Yes | NA, EU, FE – determines API endpoint |
is_active | Boolean | Yes | Whether this Amazon connection is actively processing (default: false) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
OAuth Token Refresh Sequence:
sequenceDiagram
autonumber
participant POS as POS Backend
participant CACHE as Token Cache
participant LWA as Login with Amazon
participant SPAPI as SP-API Endpoint
Note over POS, SPAPI: Token Refresh Flow
POS->>CACHE: Request access_token for tenant
alt Token Valid (expiry > now + 5 min)
CACHE-->>POS: Return cached access_token
else Token Expired or Near-Expiry
POS->>LWA: POST /auth/o2/token
Note right of LWA: grant_type=refresh_token<br/>refresh_token=***<br/>client_id=***<br/>client_secret=***
LWA-->>POS: access_token (1 hour TTL)
POS->>CACHE: Store encrypted access_token + expiry
end
POS->>SPAPI: API Request + Authorization: Bearer {access_token}
alt Success
SPAPI-->>POS: 200 OK + Response Data
else Token Rejected (401)
POS->>LWA: Force refresh token
LWA-->>POS: New access_token
POS->>SPAPI: Retry API Request
SPAPI-->>POS: 200 OK + Response Data
end
Business Rules:
- The POS system refreshes the access token proactively when the current token has less than 5 minutes remaining, avoiding mid-request expiration.
- If the refresh token itself becomes invalid (seller revokes access or Amazon rotates credentials), the integration status transitions to
DISCONNECTEDand an alert is raised to all ADMIN/OWNER users. - Each tenant may connect to at most one Amazon Seller Central account per marketplace. Multiple marketplaces within the same region are supported (e.g., US + CA + MX under NA).
- All credentials are encrypted at rest using AES-256. The
client_secret_encrypted,refresh_token_encrypted, andaccess_token_encryptedfields are write-only – API responses redact these to"***".
6.4.2 Catalog Items API
The Catalog Items API resolves Amazon Standard Identification Numbers (ASINs) from POS product identifiers, enabling product matching and listing creation.
Endpoint: GET /catalog/2022-04-01/items
Primary Use Cases:
- Look up existing ASINs by UPC/EAN barcode before creating a new listing.
- Retrieve Amazon product detail pages for enrichment (titles, bullet points, images).
- Validate that a POS product maps to the correct Amazon catalog entry.
Rate Limit: 5 requests per second, burst of 5.
Field Mapping: POS to Amazon Catalog
| POS Field | Amazon Field | Notes |
|---|---|---|
sku | seller_sku | Unique per seller account. POS SKU used as seller_sku unless overridden. |
name | item_name | Max 500 characters for Amazon. Truncated with ellipsis if POS name exceeds limit. |
long_description | product_description | HTML allowed. Max 2,000 characters. |
brand | brand | Required for most Amazon categories. Must match Amazon Brand Registry if enrolled. |
barcode | external_id (UPC/EAN) | Used for ASIN lookup. UPC-A (12 digits) or EAN-13 (13 digits). |
primary_image_url | main_image_url | Minimum 1000x1000px. Pure white background required. No watermarks or text overlays. |
base_price | price.amount | Per-marketplace pricing. Currency determined by marketplace. |
weight | item_weight | Required for FBA. Must include unit (lb, kg, oz, g). |
product_type | product_type | Amazon Browse Node taxonomy. Mapped via Amazon Product Type Definition API. |
color | color_name | Amazon standard color values. Custom colors must map to nearest Amazon standard. |
size | size_name | Amazon standard size values. Apparel uses specific size systems (US, EU, UK). |
material | material_type | Required for certain categories (apparel, jewelry). |
ASIN Resolution Workflow:
flowchart TD
A[POS Product Selected for Amazon Listing] --> B{Has UPC/EAN Barcode?}
B -->|Yes| C[Search Catalog Items API by barcode]
B -->|No| D[Search by product name + brand]
C --> E{ASIN Found?}
D --> E
E -->|Yes, Single Match| F[Auto-link ASIN to POS product]
E -->|Yes, Multiple Matches| G[Present matches to admin for selection]
E -->|No Match| H[Create new listing - ASIN assigned by Amazon]
F --> I[Store ASIN in amazon_product_mapping table]
G --> I
H --> I
I --> J[Product ready for listing creation]
Amazon Product Mapping Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
product_id | UUID | Yes | FK to POS products table |
variant_id | UUID | No | FK to POS product_variants table (null for simple products) |
marketplace_id | String(20) | Yes | Amazon marketplace (e.g., ATVPDKIKX0DER) |
asin | String(10) | No | Amazon Standard Identification Number. Null until resolved. |
seller_sku | String(40) | Yes | Seller SKU on Amazon. Defaults to POS SKU. |
fnsku | String(10) | No | Fulfillment Network SKU. Assigned by Amazon for FBA items. |
fulfillment_type | Enum | Yes | FBA, FBM, BOTH – determines fulfillment method |
listing_status | Enum | Yes | DRAFT, ACTIVE, INACTIVE, SUPPRESSED, DELETED |
last_synced_at | DateTime | No | Timestamp of most recent sync with Amazon |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
6.4.3 Listings Items API
The Listings Items API manages the creation, update, and deletion of product listings on Amazon marketplaces. POS serves as the system of record for listing data, pushing changes to Amazon.
Endpoints:
| Operation | Method | Endpoint | Use Case |
|---|---|---|---|
| Create/Update | PUT | /listings/2021-08-01/items/{sellerId}/{sellerSku} | Full listing creation or complete overwrite |
| Partial Update | PATCH | /listings/2021-08-01/items/{sellerId}/{sellerSku} | Update specific attributes only |
| Delete | DELETE | /listings/2021-08-01/items/{sellerId}/{sellerSku} | Remove listing from marketplace |
| Bulk Submit | POST | /feeds/2021-06-30/feeds | JSON_LISTINGS_FEED for bulk operations (up to 10,000 items per feed) |
Listing Lifecycle State Machine:
stateDiagram-v2
[*] --> DRAFT: Admin selects product for Amazon
DRAFT --> PENDING_REVIEW: Submit listing to Amazon
PENDING_REVIEW --> ACTIVE: Amazon approves listing
PENDING_REVIEW --> SUPPRESSED: Amazon rejects listing
ACTIVE --> INACTIVE: Admin deactivates or out of stock
ACTIVE --> SUPPRESSED: Amazon policy violation
INACTIVE --> ACTIVE: Admin reactivates with stock
SUPPRESSED --> PENDING_REVIEW: Fix issues and resubmit
ACTIVE --> DELETED: Admin permanently removes
INACTIVE --> DELETED: Admin permanently removes
DELETED --> [*]
Listing Attribute Mapping: POS to Amazon
| POS Field | Amazon Listing Attribute | Constraints | Notes |
|---|---|---|---|
name | item_name | Max 500 chars | Title formula: Brand + Product Type + Key Feature + Size/Color |
long_description | product_description | Max 2,000 chars, HTML allowed | Sanitized before push – no JavaScript or external links |
bullet_points[] | bullet_point (x5) | Max 1,000 chars each, max 5 bullets | POS stores as array. Truncated if exceeds limit. |
search_terms | generic_keyword | Max 250 bytes total | Backend keywords. No brand names, ASINs, or profanity. |
base_price | purchasable_offer.our_price | Per marketplace, currency auto-set | Converted to marketplace currency if multi-currency enabled |
compare_at_price | purchasable_offer.list_price | Must be > our_price | Used for “was/now” pricing on Amazon |
quantity | fulfillment_availability.quantity | Integer >= 0 | FBM quantity only. FBA managed by Amazon. |
condition | condition_type | new_new, used_*, refurbished | Default: new_new for retail POS |
handling_time | fulfillment_availability.handling_time | Integer (business days) | FBM only. Default: 2 days. |
Bulk Feed Processing:
For initial catalog push or large-scale updates, the system uses the Feeds API with JSON_LISTINGS_FEED type:
- POS collects up to 10,000 product listings into a single feed document.
- Feed is submitted via
POST /feeds/2021-06-30/feedswithfeedType=JSON_LISTINGS_FEED. - POS polls
GET /feeds/2021-06-30/feeds/{feedId}untilprocessingStatus=DONE. - Download the processing report to identify per-item success/failure.
- Failed items are logged to the Integration Sync Log (Module 5, Section 5.16.4) with Amazon error codes.
Business Rules:
- All required Amazon fields are validated in the POS before submission. Missing fields block the listing with a clear error message (e.g., “Brand is required for Amazon category ‘Clothing’”).
- Search terms must not contain brand names, ASINs, competitor names, or offensive language. POS applies a deny-list filter before submission.
- Bullet points are strongly recommended (Amazon penalizes listings without them in search ranking). POS displays a completeness score for Amazon-bound products.
- Price changes are applied via PATCH to avoid overwriting other listing attributes.
6.4.4 Orders API
The Orders API enables the POS system to import Amazon marketplace orders for tracking and fulfillment (FBM). FBA orders are tracked for visibility but fulfilled by Amazon.
Core Endpoints:
| Operation | Method | Endpoint | Purpose |
|---|---|---|---|
| Poll Orders | GET | /orders/v0/orders | Retrieve new/updated orders every 2 minutes |
| Get Order Items | GET | /orders/v0/orders/{orderId}/orderItems | Retrieve line item details for a specific order |
| Confirm Shipment | POST | /orders/v0/orders/{orderId}/shipment | Provide tracking for FBM fulfillment |
Order Status Mapping:
| Amazon Status | POS Status | Action |
|---|---|---|
Pending | PENDING_FULFILLMENT | Wait for Amazon payment confirmation. Do not begin fulfillment. |
Unshipped | ASSIGNED | Route to nearest fulfillment-capable store for FBM. |
PartiallyShipped | PARTIALLY_SHIPPED | Some line items shipped, remainder pending. |
Shipped | SHIPPED | Tracking provided to Amazon. Inventory decremented. |
Canceled | CANCELLED | Release reserved inventory. Log cancellation reason. |
Unfulfillable | UNFULFILLABLE | FBA cannot fulfill (e.g., out of stock at FC). Alert admin. |
InvoiceUnconfirmed | PENDING_INVOICE | EU VAT invoice required before shipment. |
PendingAvailability | BACKORDER | Pre-order. Fulfill when stock arrives. |
FBA vs FBM Fulfillment Routing:
| Fulfillment Type | Who Ships | Inventory Source | POS Action |
|---|---|---|---|
| FBA (Fulfilled by Amazon) | Amazon warehouse | Amazon Fulfillment Center (FC) | POS monitors order status only. No pick-pack-ship required. Inventory tracked as “FBA” channel. |
| FBM (Fulfilled by Merchant) | Our stores | POS physical inventory | Full pick-pack-ship workflow (Section 4.14.4). Store assignment algorithm routes to nearest location. |
| SFP (Seller Fulfilled Prime) | Our stores (Prime SLA) | POS physical inventory | Same as FBM but with Prime delivery SLA (1-2 day shipping). Requires carrier integration. |
Order Import and Fulfillment Sequence:
sequenceDiagram
autonumber
participant AMZ as Amazon SP-API
participant POLL as Order Poller (2 min)
participant API as POS Backend
participant DB as Database
participant STAFF as Store Staff
participant CARRIER as Carrier
Note over AMZ, CARRIER: Phase 1: Order Import
POLL->>AMZ: GET /orders/v0/orders?CreatedAfter={lastPoll}
AMZ-->>POLL: Order list (new + updated)
loop Each New Order
POLL->>AMZ: GET /orders/v0/orders/{orderId}/orderItems
AMZ-->>POLL: Line items with ASIN, seller_sku, qty, price
POLL->>API: Map seller_sku to POS product_id
API->>DB: Create amazon_order record
API->>DB: Create order_line_items
alt FBA Order
API->>DB: Status: TRACKING_ONLY (no fulfillment action)
else FBM Order
API->>DB: Status: PENDING_FULFILLMENT
API->>API: Run store assignment algorithm (Section 4.14.2)
API->>DB: Reserve inventory at selected store
API-->>STAFF: New Amazon FBM order on fulfillment queue
end
end
Note over AMZ, CARRIER: Phase 2: FBM Fulfillment
STAFF->>API: Begin pick-pack-ship (Section 4.14.4)
API->>DB: Status: PICKING -> PACKING -> SHIPPED
STAFF->>API: Enter carrier + tracking number
API->>AMZ: POST /orders/v0/orders/{orderId}/shipment
Note right of AMZ: carrier_code, tracking_number,<br/>ship_date, line_items
AMZ-->>API: Shipment confirmation
API->>DB: Log SALE movement, decrement inventory
Note over AMZ, CARRIER: Phase 3: Delivery Tracking
CARRIER->>API: Delivery webhook
API->>DB: Status: DELIVERED
Amazon Order Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
amazon_order_id | String(20) | Yes | Amazon order identifier (e.g., 113-1234567-1234567) |
marketplace_id | String(20) | Yes | Marketplace where order was placed |
order_status | Enum | Yes | Current Amazon order status |
pos_status | Enum | Yes | Internal POS fulfillment status |
fulfillment_channel | Enum | Yes | AFN (Amazon Fulfillment Network / FBA) or MFN (Merchant Fulfillment Network / FBM) |
assigned_location_id | UUID | No | FK to locations table. FBM orders only. |
purchase_date | DateTime | Yes | When the customer placed the order |
buyer_name | String(100) | No | Buyer display name (PII – encrypted at rest) |
shipping_address | JSONB | No | Encrypted shipping address object. Used for store assignment. |
order_total | Decimal(12,2) | Yes | Total order amount |
currency_code | String(3) | Yes | Order currency (e.g., USD) |
items_shipped | Integer | Yes | Count of shipped line items |
items_unshipped | Integer | Yes | Count of unshipped line items |
carrier_code | String(20) | No | Shipping carrier (e.g., UPS, USPS, FedEx) |
tracking_number | String(50) | No | Shipment tracking number |
shipped_at | DateTime | No | When the order was shipped |
delivered_at | DateTime | No | When the order was delivered |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Order polling runs every 2 minutes using the
CreatedAfterparameter set to the last successful poll timestamp. Pendingorders are imported for visibility but NOT queued for fulfillment until Amazon confirms payment (status transitions toUnshipped).- FBM orders reuse the existing pick-pack-ship workflow (Section 4.14.4) with Amazon-specific carrier code and tracking submission.
- Amazon buyer PII (name, address, email) is encrypted at rest and purged after 30 days per Amazon’s Acceptable Use Policy.
- If a seller_sku in an incoming order does not map to any POS product, the order is flagged as
UNMAPPEDand an alert is raised for admin resolution.
6.4.5 FBA Inventory API
The FBA Inventory API provides visibility into stock held at Amazon Fulfillment Centers. The POS system reads FBA inventory levels for unified stock reporting but does not directly control FBA quantities (Amazon manages FC inventory).
Core Endpoint: GET /fba/inventory/v1/summaries
Identifier Mapping:
| Identifier | Source | Purpose |
|---|---|---|
SKU (seller_sku) | POS system | Our internal product identifier used across all channels |
ASIN | Amazon catalog | Amazon’s product catalog identifier. One ASIN can map to multiple seller SKUs. |
FNSKU | Amazon FBA | Fulfillment Network SKU. Amazon’s internal label for FBA units. Unique per seller + product. |
UPC / EAN | Manufacturer | Universal barcode. Used for ASIN resolution during listing creation. |
MSKU | Amazon (legacy) | Merchant SKU. Equivalent to seller_sku. Used in older API versions. |
FBA Inventory States:
| State | Description | POS Tracking |
|---|---|---|
Fulfillable | Available for customer orders at Amazon FC | Shown as “FBA Available” in POS inventory view |
Inbound Working | Shipment plan created, not yet shipped to FC | Shown as “FBA Inbound” – subtotal of all inbound stages |
Inbound Shipped | In transit to FC, not yet received | Shown as “FBA Inbound” |
Inbound Receiving | Arrived at FC, being processed and stowed | Shown as “FBA Receiving” |
Reserved | Allocated to pending customer orders or FC transfers | Shown as “FBA Reserved” (not available for new orders) |
Unfulfillable | Defective, customer-damaged, carrier-damaged, or expired at FC | Triggers alert to admin – requires removal order or disposal decision |
Researching | Discrepancy under investigation by Amazon | Triggers alert to admin – monitor and follow up if unresolved > 30 days |
FBA Inventory Sync Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
product_mapping_id | UUID | Yes | FK to amazon_product_mapping table |
fnsku | String(10) | Yes | Amazon Fulfillment Network SKU |
fulfillable_qty | Integer | Yes | Units available for sale at Amazon FC |
inbound_working_qty | Integer | Yes | Units in shipment plans not yet shipped |
inbound_shipped_qty | Integer | Yes | Units in transit to FC |
inbound_receiving_qty | Integer | Yes | Units being processed at FC |
reserved_qty | Integer | Yes | Units allocated to orders or FC transfers |
unfulfillable_qty | Integer | Yes | Defective or damaged units at FC |
researching_qty | Integer | Yes | Units under investigation |
last_updated_at | DateTime | Yes | Timestamp of last data refresh from Amazon |
created_at | DateTime | Yes | Record creation timestamp |
Inbound Shipment Creation Workflow:
When the POS system needs to send inventory to Amazon FBA, it uses the Inbound Shipment API to create a shipment plan:
flowchart TD
A[Admin selects products for FBA replenishment] --> B[POS calculates quantities per SKU]
B --> C[POST createInboundShipmentPlan]
C --> D{Amazon splits into shipment groups?}
D -->|Single FC| E[One shipment plan created]
D -->|Multiple FCs| F[Multiple shipment plans created]
E --> G[Print FNSKU labels for each unit]
F --> G
G --> H[Pack and ship to assigned FC addresses]
H --> I[Enter carrier + tracking per shipment]
I --> J[POS moves inventory to FBA Inbound status]
J --> K[Poll inventory summaries for receiving confirmation]
K --> L[Amazon confirms receipt -- FBA Fulfillable updated]
Business Rules:
- FBA inventory is read-only in the POS inventory view. Staff cannot adjust FBA quantities manually – all FBA adjustments originate from Amazon.
- The POS displays a unified inventory view:
Total Available = POS On-Hand + FBA Fulfillable. Channel-specific breakdowns are shown on the product detail screen. - FBA inventory is synced every 15 minutes via
getInventorySummaries. More frequent polling is unnecessary as Amazon updates FBA quantities asynchronously. - Unfulfillable inventory exceeding a configurable threshold (default: 5 units per SKU) triggers a removal recommendation alert to the admin.
6.4.6 Notifications (SQS / EventBridge)
Amazon SP-API supports push notifications for key events, eliminating the need for constant polling. The POS system subscribes to relevant notification types and processes them via an Amazon SQS queue.
Key Notification Types:
| Notification Type | Trigger | POS Action |
|---|---|---|
ORDER_CHANGE | Order status changes (new, shipped, cancelled, returned) | Update POS order status. Trigger fulfillment queue for new FBM orders. |
ITEM_INVENTORY_EVENT_CHANGE | FBA inventory level changes (receipt, sale, adjustment) | Update FBA stock display in POS. Recalculate unified available quantity. |
LISTINGS_ITEM_STATUS_CHANGE | Listing approved, suppressed, or policy-flagged by Amazon | Update product listing_status. Alert admin for suppressed listings. |
REPORT_PROCESSING_FINISHED | A requested report (settlement, inventory) is ready for download | Download and process report. Update POS financial or inventory records. |
FBA_OUTBOUND_SHIPMENT_STATUS | FBA order shipped or delivery status updated | Update FBA order tracking info. Mark order as SHIPPED or DELIVERED. |
FEED_PROCESSING_FINISHED | A submitted feed (listings, pricing, inventory) has been processed | Download processing report. Log successes and failures per item. |
BRANDED_ITEM_CONTENT_CHANGE | A+ Content or brand content changes on a shared ASIN | Log change for awareness. No auto-action (POS does not manage A+ Content). |
SQS Queue Configuration Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
integration_id | UUID | Yes | FK to integrations table |
sqs_queue_url | String(500) | Yes | Amazon SQS queue URL for receiving notifications |
sqs_queue_arn | String(200) | Yes | SQS queue ARN for subscription registration |
aws_access_key_id_encrypted | String(200) | Yes | Encrypted AWS IAM access key for SQS polling |
aws_secret_key_encrypted | String(500) | Yes | Encrypted AWS IAM secret key |
aws_region | String(20) | Yes | AWS region where SQS queue is provisioned (e.g., us-east-1) |
subscribed_notifications | String[] | Yes | Array of notification types subscribed (e.g., ["ORDER_CHANGE", "ITEM_INVENTORY_EVENT_CHANGE"]) |
polling_interval_seconds | Integer | Yes | How often to poll SQS (default: 30 seconds) |
is_active | Boolean | Yes | Whether notification processing is enabled |
last_polled_at | DateTime | No | Timestamp of last successful SQS poll |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Notification Processing Flow:
sequenceDiagram
autonumber
participant AMZ as Amazon SP-API
participant SQS as Amazon SQS Queue
participant WORKER as POS Notification Worker
participant API as POS Backend
participant DB as Database
Note over AMZ, DB: Push Notification Flow
AMZ->>SQS: Publish notification (ORDER_CHANGE, etc.)
loop Every 30 seconds
WORKER->>SQS: ReceiveMessage (max 10 messages)
SQS-->>WORKER: Notification batch
loop Each Notification
WORKER->>WORKER: Parse notification type + payload
alt ORDER_CHANGE
WORKER->>API: Update order status
API->>DB: Write order update + log
else ITEM_INVENTORY_EVENT_CHANGE
WORKER->>API: Update FBA inventory
API->>DB: Write inventory snapshot
else LISTINGS_ITEM_STATUS_CHANGE
WORKER->>API: Update listing status
API->>DB: Write status change + alert if suppressed
else REPORT_PROCESSING_FINISHED
WORKER->>AMZ: Download completed report
WORKER->>API: Process report data
API->>DB: Write report results
end
WORKER->>SQS: DeleteMessage (acknowledge)
end
end
Business Rules:
- SQS messages are processed at-least-once. All notification handlers are idempotent – duplicate messages produce the same result without side effects.
- Failed notification processing retries up to 3 times with exponential backoff (30s, 60s, 120s). After 3 failures, the message moves to a Dead Letter Queue (DLQ) and an admin alert is raised.
- The DLQ is monitored daily. Messages in the DLQ older than 7 days are automatically logged and purged.
- When
ORDER_CHANGEnotifications are active, order polling frequency (Section 6.4.4) can be reduced from 2 minutes to 15 minutes as a fallback mechanism.
6.4.7 Rate Limits & Throttling
Amazon SP-API enforces per-endpoint rate limits using a token bucket model. The POS system must respect these limits to avoid throttling (HTTP 429) and potential API access suspension.
Per-Endpoint Rate Limits:
| Endpoint | Rate Limit | Burst | Recovery Rate | Notes |
|---|---|---|---|---|
Catalog Items (GET /catalog/...) | 5 req/sec | 5 | 1 req/sec | Shared across all catalog operations |
Listings Items (PUT/PATCH/DELETE) | 5 req/sec | 5 | 1 req/sec | Per selling partner ID |
Orders (GET /orders) | 1 req/sec | 1 | 0.5 req/sec | Shared across getOrders + getOrder |
Order Items (GET .../orderItems) | 2 req/sec | 2 | 1 req/sec | Per order ID lookup |
Feeds (POST /feeds) | 1 req/sec | 1 | 0.5 req/sec | Feed submission only |
Feed Results (GET /feeds/{id}) | 2 req/sec | 2 | 1 req/sec | Polling feed status |
Reports (POST /reports) | 1 req/sec | 1 | 0.5 req/sec | Report request submission |
| Report Download | 15 req/sec | 15 | 1 req/sec | Downloading completed reports |
FBA Inventory (GET /summaries) | 2 req/sec | 2 | 1 req/sec | Per marketplace |
Notifications (POST /subscriptions) | 1 req/sec | 1 | 0.5 req/sec | Subscription management |
Throttle Response Headers:
| Header | Purpose |
|---|---|
x-amzn-RateLimit-Limit | Current rate limit for the endpoint (requests per second) |
x-amzn-RequestId | Unique request identifier for troubleshooting with Amazon support |
Retry-After | Seconds to wait before retrying (present on 429 responses) |
Rate Limit Handling Strategy:
amazon_rate_limiting:
strategy: token_bucket_with_dynamic_adjustment
# Pre-request: check available tokens before sending
pre_request_check: true
# Track remaining capacity from response headers
header_tracking:
rate_limit_header: "x-amzn-RateLimit-Limit"
request_id_header: "x-amzn-RequestId"
# Throttle response handling (HTTP 429)
throttle_handling:
respect_retry_after: true # Wait for Retry-After seconds
fallback_wait_seconds: 30 # If no Retry-After header
max_retries: 5 # Per individual request
backoff_strategy: exponential # 1s, 2s, 4s, 8s, 16s
jitter: true # Add random 0-500ms to prevent thundering herd
# Request prioritization during high load
priority_queue:
high: # Processed first
- order_import # Customer-facing
- shipment_confirmation # Time-sensitive
- inventory_sync # Prevents overselling
medium:
- listing_updates # Can tolerate delay
- catalog_lookups # Admin-initiated
low:
- report_requests # Background tasks
- feed_status_polling # Non-urgent
# Circuit breaker: disable endpoint temporarily on repeated failures
circuit_breaker:
failure_threshold: 10 # Consecutive 429s to trigger
open_duration_seconds: 300 # Wait 5 min before retrying
half_open_requests: 2 # Test requests before fully reopening
# Utilization target (% of rate limit to use)
utilization_plans:
peak_hours: 60% # Conservative during business hours
off_peak: 90% # Aggressive during overnight sync
bulk_operations: 80% # Feed submissions and bulk updates
Business Rules:
- The POS never exceeds 90% of any endpoint’s rate limit, even during bulk operations, to leave headroom for manual admin actions.
- All API calls are routed through a centralized HTTP client that enforces rate limits before dispatch. Direct API calls bypassing the rate limiter are prohibited.
- When a 429 response is received, the system logs the event, waits the specified
Retry-Afterduration (or 30 seconds if absent), and retries with exponential backoff. - Rate limit utilization metrics are displayed on the Integration Health Dashboard (Module 5, Section 5.16.5) with per-endpoint graphs.
6.4.8 Amazon Compliance & Seller Requirements
This section defines the compliance checks, packaging standards, and business rules the POS system must enforce to maintain good standing on Amazon’s marketplace.
6.4.8.1 Seller Code of Conduct
Amazon enforces strict seller performance standards. The POS system validates data quality before pushing to Amazon to prevent listing suppressions, account warnings, and potential suspension.
Pre-Submission Validation Checklist:
| Validation | Rule | Action on Failure |
|---|---|---|
| Product title | Non-empty, <= 500 chars, no ALL CAPS, no promotional phrases (“FREE”, “SALE”) | Block listing. Show specific error. |
| Brand name | Must match Amazon Brand Registry if enrolled | Block listing. Suggest registry lookup. |
| UPC/EAN | Valid check digit, not on Amazon’s blocked barcode list | Block listing. Prompt for GS1 verification. |
| Product images | Min 1000x1000px, pure white background (RGB 255,255,255), JPEG/PNG/TIFF | Block listing. Show image requirements. |
| Price | > $0.00, within Amazon category price range, no extreme markups vs. other channels | Warning. Admin override allowed with reason code. |
| Description | Non-empty, no HTML injection, no external links, no competitor references | Sanitize and warn. Strip disallowed content. |
| Bullet points | Non-empty for at least 3 of 5 bullets, no HTML, no promotional claims | Warning. Listing proceeds but flagged as “incomplete”. |
| Condition | Valid condition type for the category | Block listing. Show valid options. |
| Weight/Dimensions | Required for FBA. Must be positive values with valid units. | Block FBA listing. Allow FBM without. |
| Restricted category | Check if product type requires Amazon approval (e.g., jewelry, pesticides) | Block listing. Provide approval application link. |
Pricing Compliance:
- Amazon monitors pricing across channels. If a product is listed significantly cheaper on the retailer’s own website (Shopify), Amazon may suppress the Buy Box or flag the listing.
- The POS displays a cross-channel price comparison alert when Amazon price differs from Shopify price by more than a configurable threshold (default: 10%).
- The system does NOT enforce price parity automatically – it alerts the admin who makes the business decision.
6.4.8.2 Packaging & Labeling Guidelines
FBA Label Requirements:
| Specification | Requirement |
|---|---|
| Label format | 4“ x 6“ thermal label (Zebra-compatible) |
| Barcode type | FNSKU barcode (Code 128 symbology) |
| Label placement | One label per sellable unit, covering any existing UPC |
| Readability | Minimum 1“ barcode height, no damage, smudging, or obstruction |
| Suffocation warning | Required on poly bags with opening > 5“ |
FBA Prep Requirements by Product Category:
| Category | Prep Required | Label Required | Special Notes |
|---|---|---|---|
| Apparel | Poly bag with suffocation warning | FNSKU barcode on bag exterior | Transparent bag preferred. No hangers for FBA. |
| Shoes | Individual shoe box | FNSKU barcode on box exterior | Each pair in separate box. No loose shoes in poly bags. |
| Accessories (belts, scarves) | Poly bag or bubble wrap | FNSKU barcode on outer packaging | Small items must be in bag to prevent loss. |
| Jewelry | Bubble wrap + rigid box | FNSKU barcode on box exterior | High-value handling. Minimum 2“ cushion padding. |
| Fragile items | Bubble wrap + rigid box | FNSKU barcode on box exterior | “FRAGILE” label optional but recommended. |
| Oversized items | None (ship as-is) | FNSKU barcode on product | Box dimensions must not exceed FBA limits. |
Shipping Box Requirements (FBA Inbound):
| Specification | Limit |
|---|---|
| Maximum box weight | 50 lbs (23 kg) |
| Maximum box dimensions | 25“ x 25“ x 25“ (standard), oversize varies |
| Box material | New corrugated cardboard, minimum 200# burst strength |
| Contents per box | Single SKU (preferred) or mixed SKU with manifest |
| Pallet shipments | Standard 40“ x 48“ pallets, max 72“ stack height |
6.4.8.3 FBA vs FBM Support
The POS system supports both Amazon fulfillment methods with per-product configuration. A single product can use FBA in one marketplace and FBM in another, or both simultaneously.
Fulfillment Method Configuration:
| Setting | Options | Default | Description |
|---|---|---|---|
fulfillment_type | FBA, FBM, BOTH | FBM | How this product is fulfilled on Amazon |
fba_location_id | UUID | null | Virtual POS location representing Amazon FBA stock (Section 6.4.8.4) |
fbm_location_ids | UUID[] | all stores | Which POS locations can fulfill FBM orders for this product |
auto_replenish_fba | Boolean | false | Whether POS auto-generates FBA inbound shipment plans when FBA stock drops below threshold |
fba_reorder_point | Integer | 0 | FBA stock level that triggers replenishment alert or auto-plan |
fba_reorder_qty | Integer | 0 | Quantity to send per FBA replenishment shipment |
FBA Shipment Creation Workflow:
sequenceDiagram
autonumber
participant ADMIN as Admin
participant POS as POS Backend
participant AMZ as Amazon SP-API
participant STORE as Store Staff
Note over ADMIN, STORE: FBA Inbound Shipment Creation
ADMIN->>POS: Select products + quantities for FBA replenishment
POS->>POS: Validate stock available at source location
POS->>AMZ: POST createInboundShipmentPlan
Note right of AMZ: SKUs, quantities,<br/>ship-from address,<br/>label preference
AMZ-->>POS: Shipment plan(s) with FC destination(s)
Note left of POS: Amazon may split into<br/>multiple shipments to<br/>different FCs
loop Each Shipment Plan
POS->>POS: Generate FNSKU labels (4x6 thermal)
POS-->>ADMIN: Display FC destination + packing instructions
ADMIN->>STORE: Assign packing task to store staff
STORE->>STORE: Apply FNSKU labels, pack boxes
STORE->>POS: Confirm shipment packed + carrier + tracking
POS->>AMZ: PUT updateInboundShipment (carrier, tracking, box contents)
POS->>POS: Move inventory: POS On-Hand -> FBA Inbound
end
Note over ADMIN, STORE: Amazon receives and processes
AMZ-->>POS: Notification: ITEM_INVENTORY_EVENT_CHANGE
POS->>POS: Move inventory: FBA Inbound -> FBA Fulfillable
6.4.8.4 Safety Buffer Rules
To prevent overselling when sync delays exist between the POS and Amazon, configurable safety buffers reduce the quantity advertised on Amazon.
Buffer Calculation:
Amazon Available = POS Allocatable Quantity - Safety Buffer
Where:
- POS Allocatable Quantity = On-hand at designated Amazon fulfillment location(s), minus reserved, minus safety stock.
- Safety Buffer = Greater of (percentage buffer) or (minimum unit buffer).
Buffer Configuration:
| Setting | Type | Default | Description |
|---|---|---|---|
buffer_percentage | Decimal | 10% | Percentage of allocatable quantity to withhold |
buffer_minimum_units | Integer | 2 | Minimum units to always withhold, regardless of percentage |
buffer_scope | Enum | PER_PRODUCT | PER_PRODUCT (individual) or GLOBAL (all products same rule) |
amazon_location_id | UUID | null | Dedicated virtual location for Amazon inventory (recommended) |
use_specific_location | Boolean | false | If true, only inventory at amazon_location_id is available for Amazon. If false, network-wide available. |
Examples:
| Scenario | POS Available | Buffer % | Min Units | Buffer Applied | Amazon Qty |
|---|---|---|---|---|---|
| Normal stock | 50 | 10% | 2 | 5 (10% of 50) | 45 |
| Low stock | 10 | 10% | 2 | 2 (min units > 10% of 10) | 8 |
| Very low stock | 3 | 10% | 2 | 2 (min units) | 1 |
| Minimal stock | 2 | 10% | 2 | 2 (min units) | 0 |
| Out of stock | 0 | 10% | 2 | 0 | 0 |
Business Rules:
- When
Amazon Availabledrops to 0, the listing quantity is set to 0 on Amazon. The listing remains ACTIVE but shows “Currently unavailable.” - Safety buffers are recalculated on every inventory change event (sale, receiving, adjustment, transfer).
- Admins can override the buffer for individual products (e.g., set buffer to 0 for high-velocity items where overselling risk is acceptable).
- The dedicated Amazon virtual location approach is recommended for retailers with significant Amazon volume. It provides a clear physical segregation of Amazon-allocated inventory.
6.4.8.5 Order Routing Rules
Amazon FBM orders are integrated into the existing fulfillment infrastructure alongside Shopify orders.
Routing Logic:
flowchart TD
A[Amazon Order Received] --> B{Fulfillment Channel?}
B -->|AFN / FBA| C[Track status only]
C --> D[Display in POS as FBA order]
D --> E[Monitor via ITEM_INVENTORY_EVENT_CHANGE]
B -->|MFN / FBM| F{SFP / Prime?}
F -->|Prime SLA| G[Flag as HIGH PRIORITY]
F -->|Standard FBM| H[Standard priority]
G --> I[Run store assignment algorithm]
H --> I
I --> J{Stock available at nearest store?}
J -->|Yes| K[Assign to nearest store]
J -->|No| L{Stock available at ANY store?}
L -->|Yes| M[Assign to store with stock - warn about shipping cost]
L -->|No| N{Split fulfillment enabled?}
N -->|Yes| O[Split across multiple stores]
N -->|No| P[Alert admin: cannot fulfill]
K --> Q[Add to store fulfillment queue]
M --> Q
O --> Q
Q --> R[Pick-Pack-Ship workflow<br/>Section 4.14.4]
R --> S[POST shipment confirmation to Amazon]
Priority Rules for Mixed Channel Fulfillment:
| Priority | Source | SLA | Notes |
|---|---|---|---|
| 1 (Highest) | Amazon Prime (SFP) | 1-2 business days | Must ship same day if ordered before cutoff |
| 2 | Amazon FBM Standard | 3-5 business days | Ship within handling_time (default: 2 days) |
| 3 | Shopify orders | Per shipping method selected | Existing Shopify fulfillment SLA applies |
| 4 | In-store pickup | Customer-defined | Lower urgency, customer comes to store |
Business Rules:
- Amazon FBM orders appear in the same fulfillment queue as Shopify orders, with priority badges indicating the channel and SLA.
- The store assignment algorithm (Section 4.14.2) is reused for Amazon FBM orders with the same proximity + stock availability logic.
- If an FBM order cannot be fulfilled within the stated handling time, the system alerts the admin to either fulfill from an alternate location or request a cancellation from the buyer.
- Late shipments are tracked as a seller performance metric. Amazon penalizes sellers with >4% late shipment rate.
6.4.9 Reports: Amazon Integration
All Amazon integration reports are accessible from the Admin Portal reporting module with date range filtering, export to CSV/PDF, and drill-down capability.
| Report | Purpose | Key Data Fields |
|---|---|---|
| Amazon Sales Summary | Daily/weekly/monthly sales performance across Amazon channel | Date range, total revenue, order count, units sold, avg order value, FBA vs FBM breakdown, refund rate, marketplace breakdown |
| Amazon Inventory Status | Current stock levels across FBA and FBM channels | SKU, product name, FBA fulfillable qty, FBA inbound qty, FBA reserved qty, FBA unfulfillable qty, FBM available qty, total Amazon available |
| Amazon Listing Health | Status of all Amazon product listings | SKU, ASIN, listing status (active/inactive/suppressed), suppression reason, Buy Box %, listing completeness score, last sync timestamp |
| Amazon Order Fulfillment | Fulfillment performance metrics | Order count, avg fulfillment time (hours), on-time shipment %, late shipment %, carrier breakdown, FBA vs FBM split, return rate |
| Amazon Fee Analysis | Total Amazon costs and profitability per SKU | SKU, ASIN, referral fee %, referral fee $, FBA fee (pick + pack + weight), storage fee (monthly + long-term), total Amazon fees, POS cost, net margin after fees |
| Amazon Sync Health | Integration connectivity and sync reliability | Sync success rate (24h), failed syncs (with error codes), avg sync latency, API quota utilization %, notification delivery rate, DLQ message count |
| Amazon Compliance Scorecard | Seller account health metrics from Amazon | Order defect rate (target <1%), late shipment rate (target <4%), pre-fulfillment cancel rate (target <2.5%), valid tracking rate (target >95%), policy violations |
Cross-Channel Comparison View:
A dedicated cross-channel report enables the admin to compare performance across POS in-store, Shopify online, and Amazon channels:
| Metric | In-Store (POS) | Shopify Online | Amazon FBM | Amazon FBA |
|---|---|---|---|---|
| Revenue | Sum of in-store sales | Sum of Shopify orders | Sum of FBM order revenue | Sum of FBA order revenue |
| Units Sold | POS transaction items | Shopify line items | FBM shipped units | FBA shipped units |
| Avg Order Value | POS avg | Shopify avg | FBM avg | FBA avg |
| Margin % | After POS costs | After Shopify fees | After Amazon referral + shipping | After Amazon referral + FBA fees |
| Return Rate | In-store returns | Shopify returns | FBM returns | FBA returns |
Business Rules:
- Amazon fee data is sourced from Amazon Settlement Reports, requested bi-weekly and reconciled against POS order records.
- The Amazon Compliance Scorecard polls Amazon’s Seller Performance API daily and surfaces warnings when any metric approaches the threshold (yellow at 75% of limit, red at 90%).
- All reports respect the tenant’s timezone for date boundaries and the tenant’s currency for financial values.
- Report data is cached for 1 hour. An admin can force-refresh any report to pull the latest data from Amazon.
6.4.10 Amazon Integration Configuration (YAML Reference)
The following YAML reference consolidates all Amazon integration business rules for the rules engine:
amazon_integration:
version: "1.0"
# Connection settings
connection:
oauth_token_refresh_buffer_minutes: 5
max_marketplaces_per_tenant: 3
credential_encryption: AES-256
pii_retention_days: 30
# Sync intervals
sync_intervals:
order_polling_minutes: 2
order_polling_with_notifications_minutes: 15
fba_inventory_sync_minutes: 15
listing_sync_on_change: true
daily_reconciliation_hour: 3 # 3 AM tenant-local
# Safety buffer defaults
safety_buffer:
default_percentage: 10
default_minimum_units: 2
scope: PER_PRODUCT
recalculate_on: [SALE, RECEIVING, ADJUSTMENT, TRANSFER, FBA_UPDATE]
# Order routing
order_routing:
fbm_use_store_assignment_algorithm: true
prime_same_day_cutoff_hour: 14 # 2 PM local
default_handling_time_days: 2
late_shipment_alert_threshold_hours: 36
# Notification processing
notifications:
sqs_polling_interval_seconds: 30
max_messages_per_poll: 10
retry_max_attempts: 3
retry_backoff: [30, 60, 120] # seconds
dlq_retention_days: 7
# Compliance thresholds (Amazon seller performance)
compliance:
order_defect_rate_warning: 0.75 # Yellow at 0.75%
order_defect_rate_critical: 0.90 # Red at 0.90% (limit: 1%)
late_shipment_rate_warning: 3.0 # Yellow at 3%
late_shipment_rate_critical: 3.6 # Red at 3.6% (limit: 4%)
cancel_rate_warning: 1.875 # Yellow at 1.875%
cancel_rate_critical: 2.25 # Red at 2.25% (limit: 2.5%)
valid_tracking_warning: 96.25 # Yellow below 96.25%
valid_tracking_critical: 95.0 # Red below 95% (limit: 95%)
# Listing validation
listing_validation:
title_max_chars: 500
description_max_chars: 2000
bullet_points_max: 5
bullet_point_max_chars: 1000
search_terms_max_bytes: 250
image_min_dimension_px: 1000
price_cross_channel_alert_threshold_percent: 10
# FBA settings
fba:
unfulfillable_alert_threshold_units: 5
researching_followup_days: 30
label_format: "4x6_thermal"
label_barcode_symbology: "CODE128"
max_box_weight_lbs: 50
max_box_dimension_inches: 25
End of Section 6.4: Amazon SP-API Integration
6.5 Google Merchant API Integration
Scope: Outbound product-feed management, local inventory advertising, Google Business Profile linkage, push notification handling, and Content API-to-Merchant API migration for Google Shopping and Local Inventory Ads.
Cross-Reference: See Module 3, Section 3.9 for the product-sync lifecycle that triggers outbound calls to Google Merchant Center.
Cross-Reference: See Module 4, Section 4.5 for inventory-level sync events that Module 6 pushes to connected channels including Google local inventory.
Cross-Reference: See Module 5, Section 5.4 for tenant-level integration configuration (credentials, enabled providers, sync schedules).
CRITICAL: The Content API for Shopping reaches end-of-life on August 18, 2026. All new development MUST target the Merchant API (v1beta / v1). See Section 6.5.10 for the migration plan.
6.5.1 Authentication & Authorization
Google Merchant Center uses OAuth 2.0 with service accounts for server-to-server authentication. Unlike Shopify’s long-lived access tokens, Google service accounts issue short-lived JWTs that must be self-signed and exchanged for access tokens every 60 minutes.
Service Account Auth Flow
sequenceDiagram
autonumber
participant POS as POS Backend
participant VAULT as Credential Vault
participant JWT as JWT Builder
participant GAUTH as Google OAuth2 Endpoint
participant GAPI as Google Merchant API
Note over POS, GAPI: Phase 1: Credential Retrieval
POS->>VAULT: Fetch service account key (AES-256 decrypted)
VAULT-->>POS: Return private_key, client_email, project_id
Note over POS, GAPI: Phase 2: JWT Assertion
POS->>JWT: Build self-signed JWT
Note right of JWT: iss = client_email<br/>scope = content, merchantapi<br/>aud = https://oauth2.googleapis.com/token<br/>iat = now, exp = now + 3600
JWT->>JWT: Sign with RSA-SHA256 (private_key)
JWT-->>POS: Return signed JWT assertion
Note over POS, GAPI: Phase 3: Token Exchange
POS->>GAUTH: POST /token (grant_type=jwt-bearer, assertion=JWT)
GAUTH-->>POS: Return access_token (expires_in: 3600)
POS->>POS: Cache access_token (refresh at T-5min)
Note over POS, GAPI: Phase 4: API Call
POS->>GAPI: GET /accounts/{merchantId}/products<br/>Authorization: Bearer {access_token}
GAPI-->>POS: 200 OK (product data)
Note over POS, GAPI: Phase 5: Auto-Refresh (background)
POS->>POS: Timer fires at T-5min before expiry
POS->>GAUTH: POST /token (new JWT assertion)
GAUTH-->>POS: Return fresh access_token
Merchant Center Account Setup
Before API access is possible, the tenant must:
- Create a Merchant Center account at https://merchants.google.com
- Create a Google Cloud project and enable the Merchant API
- Create a service account in the Cloud Console with the
Merchant Center Adminrole - Download the JSON key file and upload it to the POS Admin Portal (Module 5)
- Link the service account to the Merchant Center account via the MC settings panel
- Verify and claim the website URL associated with the product feed
Data Model: google_merchant_credentials
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants.id |
merchant_account_id | VARCHAR(20) | Yes | Google Merchant Center account ID (numeric) |
service_account_email | VARCHAR(255) | Yes | Service account email (e.g., pos-sync@project.iam.gserviceaccount.com) |
private_key_encrypted | TEXT | Yes | RSA private key, AES-256-GCM encrypted at rest |
project_id | VARCHAR(100) | Yes | Google Cloud project ID |
token_endpoint | VARCHAR(255) | Yes | Default: https://oauth2.googleapis.com/token |
access_token | TEXT | No | Current cached access token (encrypted) |
token_expiry | TIMESTAMPTZ | No | Expiry timestamp of current access token |
scopes | TEXT[] | Yes | Granted scopes: content, merchantapi |
is_active | BOOLEAN | Yes | Whether this credential set is enabled – default true |
last_auth_at | TIMESTAMPTZ | No | Timestamp of last successful authentication |
last_auth_error | TEXT | No | Last authentication error message (if any) |
created_at | TIMESTAMPTZ | Yes | Row creation timestamp |
updated_at | TIMESTAMPTZ | Yes | Last modification timestamp |
Security Note: The
private_key_encryptedfield uses AES-256-GCM with a tenant-specific key derived from the master encryption key. The plaintext private key is never written to logs, error messages, or API responses. See Module 5, Section 5.4.5 for the credential vault architecture.
6.5.2 Product Data Management
The Google Merchant API separates product writes and reads into distinct resources:
ProductInput– the write resource. Used to create or update product data. Endpoint:accounts/{account}/productInputs:insertProduct– the read-only processed resource. Represents the product after Google’s validation, enrichment, and policy review. Endpoint:accounts/{account}/products/{product}
This distinction matters because the data you submit (ProductInput) may differ from the data Google stores (Product) after processing. The POS must read the Product resource to check approval status, supplemental attributes, and policy violations.
API Endpoint Reference
| Operation | Merchant API Endpoint | HTTP Method | Notes |
|---|---|---|---|
| Create/update product | accounts/{account}/productInputs:insert | POST | Upsert by offerId + feedLabel + contentLanguage |
| Get processed product | accounts/{account}/products/{product} | GET | Returns Google-enriched version with status |
| List products | accounts/{account}/products | GET | Paginated, max 250 per page |
| Delete product input | accounts/{account}/productInputs/{productInput}:delete | DELETE | Removes from feed; may take up to 24h to delist |
| Get product status | accounts/{account}/productStatuses/{product} | GET | Approval status, disapproval reasons, warnings |
Content API End-of-Life
| Milestone | Date | Impact |
|---|---|---|
| Merchant API v1beta GA | Available now | New features only in Merchant API |
| Content API deprecation announcement | 2025 | No new features; maintenance only |
| v1beta migration deadline | February 28, 2026 | Must begin migration |
| Content API full EOL | August 18, 2026 | All Content API endpoints cease functioning |
CRITICAL: Any integration built against the Content API (
/content/v2.1/) will stop working on August 18, 2026. The POS MUST use the Merchant API from day one. See Section 6.5.10 for legacy migration details.
Product Field Mapping: POS to Google Merchant
| POS Field | Google Field | API Path | Notes |
|---|---|---|---|
sku | offerId | productInput.offerId | Max 50 chars, unique per feed. Used as dedup key. |
name | title | productInput.title | Max 150 chars. No promotional text (“Sale!”, “Free Shipping” prohibited). |
long_description | description | productInput.description | Max 5,000 chars. Plain text preferred; HTML tags stripped by Google. |
primary_image_url | imageLink | productInput.imageLink | HTTPS required. Min 100x100px (apparel: 250x250px). |
base_price + currency | price | productInput.price | Object: { amountMicros: 2999000000, currencyCode: "USD" }. Price in micros (1 USD = 1,000,000 micros). |
| Computed from inventory | availability | productInput.availability | Enum: in_stock, out_of_stock, preorder, backorder. Derived from real-time inventory qty. |
brand | brand | productInput.brand | Required for most categories. Must be manufacturer brand, not store name. |
barcode (UPC/EAN) | gtin | productInput.gtin | Valid GTIN-8/12/13/14. No dashes. Check digit validated (mod-10 algorithm). |
product_condition | condition | productInput.condition | Enum: new, refurbished, used. Default: new. |
website_url | link | productInput.link | Must be live, HTTPS, product data on page must match feed exactly. |
manufacturer_part_number | mpn | productInput.mpn | Required if no GTIN available. Manufacturer’s part number. |
product_type_taxonomy | googleProductCategory | productInput.googleProductCategory | Full Google taxonomy path (e.g., “Apparel & Accessories > Clothing > Shirts”). |
weight + weight_unit | shippingWeight | productInput.shippingWeight | Object: { value: 0.5, unit: "lb" }. |
additional_images[] | additionalImageLinks | productInput.additionalImageLinks | Up to 10 additional images. Same quality requirements as primary. |
variant_color | color | productInput.color | Required for apparel. Google-normalised colour names. |
variant_size | size | productInput.size | Required for apparel. Use standard sizing (S, M, L, XL or numeric). |
Price Format Conversion
The POS stores prices as DECIMAL(10,2) but Google requires prices in micros (millionths of the currency unit):
POS price: $29.99 (DECIMAL)
Google micros: 29990000 (INT64)
Conversion: price_micros = ROUND(pos_price * 1_000_000)
6.5.3 Local Inventory
Local Inventory Ads (LIA) allow the POS to surface real-time store-level availability directly in Google Shopping results. When a shopper searches for a product near a physical store, Google can display “In stock at [Store Name]” with options for in-store pickup or same-day delivery.
Local Inventory API
| Operation | Endpoint | Method | Notes |
|---|---|---|---|
| Insert local inventory | accounts/{account}/products/{product}/localInventories:insert | POST | Upsert by storeCode |
| List local inventories | accounts/{account}/products/{product}/localInventories | GET | Returns all store-level records for a product |
| Delete local inventory | accounts/{account}/products/{product}/localInventories/{storeCode}:delete | DELETE | Remove store-level entry |
Processing time: Updates take up to 30 minutes to appear in Google Shopping results. The POS should not expect real-time reflection.
POS Location to Google Store Code Mapping
Each POS location must be mapped to a Google storeCode (the unique identifier from Google Business Profile). This mapping is configured in Module 5 and stored in the google_store_mappings table.
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants.id |
pos_location_id | UUID | Yes | FK to locations.id (Module 5) |
google_store_code | VARCHAR(50) | Yes | Google Business Profile store code |
google_merchant_account_id | VARCHAR(20) | Yes | FK reference to google_merchant_credentials.merchant_account_id |
store_name | VARCHAR(200) | Yes | Display name for the store |
is_lia_enrolled | BOOLEAN | Yes | Whether Local Inventory Ads are enabled for this location – default false |
pickup_method | VARCHAR(30) | No | Default: buy. Options: buy, reserve, ship_to_store, not_supported |
pickup_sla | VARCHAR(20) | No | Default: same_day. Options: same_day, next_day, multi_day, multi_week |
is_active | BOOLEAN | Yes | Whether this mapping is active – default true |
last_sync_at | TIMESTAMPTZ | No | Timestamp of most recent local inventory sync |
created_at | TIMESTAMPTZ | Yes | Row creation timestamp |
updated_at | TIMESTAMPTZ | Yes | Last modification timestamp |
Local Inventory Data Payload
For each product-location combination, the POS pushes:
| Field | Type | Required | Description |
|---|---|---|---|
storeCode | STRING | Yes | Mapped from google_store_mappings.google_store_code |
availability | STRING | Yes | in_stock, out_of_stock, limited_availability |
price | PRICE | No | Store-specific price override (if different from online). Object: { amountMicros, currencyCode } |
salePrice | PRICE | No | Store-specific sale price. Object: { amountMicros, currencyCode } |
salePriceEffectiveDate | STRING | No | ISO 8601 interval: 2026-03-01T00:00:00Z/2026-03-15T23:59:59Z |
quantity | INT64 | No | Exact quantity in stock at this location. Google recommends providing this. |
pickupMethod | STRING | No | buy, reserve, ship_to_store, not_supported |
pickupSla | STRING | No | same_day, next_day, multi_day, multi_week |
Local Inventory Sync Architecture
flowchart TD
INV_EVENT[Inventory Change Event<br/>Module 4] -->|qty changed| SYNC_EVAL{Sync Evaluation}
SYNC_EVAL -->|Google Merchant enabled<br/>& location enrolled| BUILD[Build Local Inventory Payload]
SYNC_EVAL -->|Not enrolled or<br/>provider disabled| SKIP[Skip -- No Action]
BUILD --> MAP[Map POS location_id<br/>to Google storeCode]
MAP --> AVAIL{Compute Availability}
AVAIL -->|qty > threshold| IN_STOCK[availability = in_stock]
AVAIL -->|qty > 0 AND<br/>qty <= threshold| LIMITED[availability = limited_availability]
AVAIL -->|qty == 0| OOS[availability = out_of_stock]
IN_STOCK --> BATCH
LIMITED --> BATCH
OOS --> BATCH
BATCH[Batch Queue<br/>Max 10 per request] -->|Flush every 60s<br/>or batch full| GAPI[POST localInventories:insert]
GAPI -->|200 OK| LOG_OK[Log Success<br/>Update last_sync_at]
GAPI -->|4xx / 5xx| RETRY[Retry Pipeline<br/>Section 6.2.3]
RETRY -->|Exhausted| DLQ[Dead Letter Queue]
Cross-Reference: See Module 4, Section 4.5.2 for the inventory change event schema that triggers local inventory sync.
6.5.4 Push Notifications
Google Merchant Center can send push notifications to a registered HTTPS callback endpoint when product statuses change. This eliminates the need to poll the Product Status API repeatedly.
Notification Configuration
| Parameter | Requirement |
|---|---|
| Callback URL | Must be publicly accessible HTTPS endpoint |
| SSL Certificate | Must be signed by a trusted CA (no self-signed certificates) |
| Response time | Must respond with 200 OK within 10 seconds |
| Authentication | Google sends a JWT in the Authorization header; validate against Google public keys |
Supported Notification Types
| Notification Type | Trigger | Payload Fields |
|---|---|---|
PRODUCT_STATUS_CHANGE | Product approval status changes (approved, disapproved, pending) | productId, accountId, attribute, previousValue, newValue, timestamp |
ACCOUNT_STATUS_CHANGE | Merchant Center account status changes | accountId, attribute, previousValue, newValue, timestamp |
Callback Data Model: google_merchant_notifications
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants.id |
notification_id | VARCHAR(100) | Yes | Google-provided unique notification ID (idempotency key) |
notification_type | VARCHAR(50) | Yes | PRODUCT_STATUS_CHANGE, ACCOUNT_STATUS_CHANGE |
product_id | VARCHAR(100) | No | Google product ID (present for product notifications) |
attribute | VARCHAR(100) | Yes | The attribute that changed (e.g., status, approvalStatus) |
previous_value | TEXT | No | Value before the change |
new_value | TEXT | Yes | Value after the change |
raw_payload | JSONB | Yes | Complete raw notification payload for audit |
processed | BOOLEAN | Yes | Whether the notification has been handled – default false |
processed_at | TIMESTAMPTZ | No | When the notification was processed |
received_at | TIMESTAMPTZ | Yes | When the POS received the notification |
Notification Processing Flow
When a PRODUCT_STATUS_CHANGE notification indicates a disapproval:
- Parse the notification payload and extract the
productIdand disapproval reason. - Map the Google
productIdback to the POSskuusing theofferIdfield. - Update the product’s Google sync status in the POS database.
- If the disapproval reason is actionable (e.g., missing GTIN, low image quality), create an alert for the catalog team (Module 3).
- Log the notification in
google_merchant_notificationsfor audit.
6.5.5 Rate Limits
Google Merchant API enforces rate limits at multiple levels. The POS must respect these limits to avoid request rejection and potential account suspension.
Rate Limits by Operation
| Operation | Limit | Strategy |
|---|---|---|
Product insert/update (productInputs:insert) | 2 updates per product per day | Batch all changes for a product; push at most 2x daily. Use change-detection to skip unchanged products. |
Product get (products.get) | Generous (1,000+ per day) | Cache responses locally for 15 minutes. Use bulk products.list instead of individual gets. |
Product list (products.list) | 250 results per page | Implement cursor-based pagination. Store nextPageToken for continuation. |
Local inventory insert (localInventories:insert) | No strict per-product limit | Update on inventory change events. Batch up to 10 entries per request. |
| Account-level daily quota | Varies by account tier | Monitor X-RateLimit-Remaining headers. Schedule bulk syncs during 02:00-06:00 UTC off-peak window. |
| Requests per minute | ~600 for standard accounts | Implement in-memory token bucket per tenant. Queue excess requests. |
Batching Strategy
To stay within the 2-updates-per-product-per-day limit:
- Change Detection: Before syncing, compare the current product data hash against the last-synced hash. Skip unchanged products.
- Batch Window: Accumulate product changes during the business day. Execute sync at two scheduled windows (e.g., 06:00 and 18:00 UTC).
- Priority Queue: If a product is updated more than twice in a day, only the most recent state is synced. Intermediate states are discarded.
- Emergency Override: Critical changes (price corrections, safety-related updates) bypass the batch window and sync immediately, consuming one of the two daily slots.
google_merchant_sync:
batch_windows:
- "06:00 UTC"
- "18:00 UTC"
max_updates_per_product_per_day: 2
change_detection: sha256_hash
emergency_override_enabled: true
emergency_reasons:
- price_correction
- safety_recall
- legal_compliance
off_peak_bulk_window: "02:00-06:00 UTC"
6.5.6 Google Required Product Data Fields
Google Shopping enforces strict data requirements. Products missing required fields or containing policy-violating content will be disapproved and will not appear in Shopping results.
Mandatory Feed Fields – POS Must Provide
| Google Field | POS Source Field | Required | Validation Rule |
|---|---|---|---|
id (offerId) | sku | Yes | Unique per product, max 50 chars, alphanumeric + hyphens + underscores only |
title | name | Yes | Max 150 chars. Prohibited: “Sale!”, “Free Shipping”, “Best Price”, all-caps words, excessive punctuation |
description | long_description | Yes | Max 5,000 chars. Plain text preferred. No HTML tags (Google strips them). No promotional language. |
imageLink | primary_image_url | Yes | Min 100x100px (apparel: 250x250px). No watermarks, no promotional overlays, no placeholder images. Must be publicly accessible HTTPS URL. |
price | base_price + currency_code | Yes | Must match website/landing page price EXACTLY. Object: { amountMicros, currencyCode }. Price discrepancies cause immediate disapproval. |
availability | Computed from inventory | Yes | Enum: in_stock, out_of_stock, preorder, backorder. Must match actual stock. Mismatches cause disapproval. |
brand | brand | Yes (most categories) | Manufacturer brand name, NOT store name. Required for all products with a known brand. |
gtin | barcode (UPC/EAN) | Yes (if exists) | Valid GTIN-8/12/13/14. No dashes or spaces. Check digit validated via mod-10 algorithm. |
condition | product_condition | Yes | Enum: new, refurbished, used. Default: new. Must accurately reflect product condition. |
link | Website product URL | Yes | Must resolve to a live HTTPS page. Product data on page must match feed data. Broken links cause disapproval. |
mpn | manufacturer_part_number | Conditional | Required if no GTIN available. Must be the manufacturer’s own part number. |
Optional but Recommended Fields
| Google Field | POS Source Field | Benefit |
|---|---|---|
additionalImageLink | additional_images[] (up to 10) | Better product presentation; higher click-through rate |
productHighlight | bullet_points[] (up to 10, max 150 chars each) | Enhanced listing appearance in Google Shopping |
color | variant_color | Required for apparel; improves search matching |
size | variant_size | Required for apparel; enables size-based filtering |
material | material_type | Improves product matching and filtering accuracy |
pattern | pattern_type | Improves matching for patterned products (striped, plaid, etc.) |
ageGroup | target_age_group | Required for apparel. Enum: newborn, infant, toddler, kids, adult |
gender | target_gender | Required for apparel. Enum: male, female, unisex |
itemGroupId | parent_product_id | Groups variants together in Shopping results |
salePrice | sale_price + currency_code | Displays original + sale price; strikethrough pricing |
salePriceEffectiveDate | sale_start / sale_end | ISO 8601 interval for automatic sale price activation |
shippingWeight | weight + weight_unit | Required for carrier-calculated shipping rates |
6.5.7 Google Image Requirements
Product images are the single most important factor in Google Shopping performance. Google enforces strict image quality rules; violations result in product disapproval.
Image Specification Table
| Requirement | Specification | POS Validation |
|---|---|---|
| Minimum size | 100x100 pixels (apparel: 250x250px) | Validate dimensions on upload; block submission if below minimum |
| Maximum size | 64 megapixels / 16 MB file size | Validate on upload; reject files exceeding limits |
| Format | JPEG, PNG, GIF (non-animated), BMP, TIFF, WebP | Validate file extension AND MIME type (prevent extension spoofing) |
| Background | White or transparent preferred | Advisory warning if background is non-white; do not block |
| Watermarks | PROHIBITED – causes immediate disapproval | Block upload if watermark metadata detected; warn if image analysis flags overlay text |
| Promotional text | PROHIBITED – no “Sale”, “Free Shipping”, logos, badges | Advisory warning to user; flag for manual review before sync |
| Borders | No decorative borders allowed | Advisory warning if border detected in image analysis |
| Product visibility | Product must occupy 75-90% of image frame | Advisory warning based on object-detection analysis (if available) |
| URL accessibility | Must be publicly accessible via HTTPS | Validate URL reachability (HTTP HEAD request) before every sync |
| Image alt text | Descriptive, non-keyword-stuffed | Auto-generate from product name + colour + size if empty |
| Placeholders | No “image coming soon” or generic placeholder images | Block sync if image URL matches known placeholder patterns |
| Product accuracy | Image must show the exact product being sold | Manual review flag; cannot be automated reliably |
Image Validation Workflow
The POS validates images at two stages:
- Upload time (immediate feedback): File size, dimensions, format, MIME type.
- Pre-sync time (before pushing to Google): URL accessibility, watermark detection, placeholder detection, completeness check.
6.5.8 Google Disapproval Prevention Rules
Product disapprovals directly reduce revenue by removing products from Google Shopping. The POS must implement proactive validation to prevent disapprovals before they occur.
Common Disapproval Reasons & POS Prevention
| Disapproval Reason | Google Policy | POS Prevention Rule |
|---|---|---|
| Price mismatch | Feed price MUST match landing page price exactly | Cross-check feed price vs. website price before every sync. Block sync if prices diverge by > $0.01. |
| Availability mismatch | Feed availability MUST match actual stock | Real-time inventory sync. Automatically set out_of_stock when qty == 0. Never show in_stock for zero-qty items. |
| Missing required attributes | All mandatory fields must be populated | Pre-sync validation gate blocks submission if any required field is empty or null. |
| Low image quality | Below minimum resolution, watermarks, promotional overlays | Image validation on upload (dimensions, format). Pre-sync URL accessibility check. |
| Misrepresentation | No fake urgency, fake scarcity, misleading claims | Block promotional text patterns in title and description: regex filter for “Sale!”, “Limited Time”, “Act Now”, etc. |
| Prohibited content | Restricted product categories (weapons, drugs, counterfeit) | Flag restricted Google product categories during product setup. Require manager approval before sync. |
| Missing landing page | Product URL must resolve to a live page with matching data | HTTP HEAD request to link URL before every sync. Block sync on 404/500 responses. |
| Invalid GTIN | Wrong format, invalid check digit, or mismatched product | GTIN validation algorithm: verify length (8/12/13/14 digits), compute mod-10 check digit, reject on mismatch. |
| Missing business info | Merchant Center must have contact info, return policy | Validate Merchant Center profile completeness via API during initial setup. Alert if incomplete. |
| Insufficient product identifiers | Products with known UPC must include gtin | Require barcode field when product has_manufacturer_upc = true. Warn if GTIN is empty for branded products. |
POS Pre-Sync Validation Checklist (Automated)
Every product must pass all 10 validation checks before being submitted to Google Merchant:
- All required fields populated (non-null, non-empty)?
- Primary image meets quality requirements (dimensions, format, HTTPS)?
- Price matches across all channels (POS, website, feed)?
- GTIN/UPC valid format (correct length, check digit passes mod-10)?
- Product category not in prohibited list?
- Landing page URL accessible (HTTP 200 response)?
- Title length <= 150 chars, no promotional text detected?
- Description length <= 5,000 chars, no HTML tags?
- Brand field populated (non-empty, not equal to store name)?
- Availability computed from current inventory (not stale > 1 hour)?
Pre-Sync Validation Flowchart
flowchart TD
START[Product Queued for<br/>Google Sync] --> V1{1. Required fields<br/>populated?}
V1 -->|No| FAIL_REQ[BLOCK: Missing required<br/>fields -- log which fields]
V1 -->|Yes| V2{2. Image meets<br/>quality requirements?}
V2 -->|No| FAIL_IMG[BLOCK: Image quality<br/>failure -- log reason]
V2 -->|Yes| V3{3. Price matches<br/>across channels?}
V3 -->|No| FAIL_PRICE[BLOCK: Price mismatch<br/>detected -- log discrepancy]
V3 -->|Yes| V4{4. GTIN valid<br/>format & check digit?}
V4 -->|No GTIN & no MPN| FAIL_ID[BLOCK: Missing product<br/>identifier -- GTIN or MPN required]
V4 -->|Invalid check digit| FAIL_GTIN[BLOCK: Invalid GTIN<br/>check digit -- log expected vs actual]
V4 -->|Valid or N/A| V5{5. Category not<br/>prohibited?}
V5 -->|Prohibited| FAIL_CAT[BLOCK: Prohibited<br/>category -- requires override]
V5 -->|Allowed| V6{6. Landing page<br/>URL accessible?}
V6 -->|404 / 500 / timeout| FAIL_URL[BLOCK: Landing page<br/>not accessible -- log HTTP status]
V6 -->|200 OK| V7{7. Title ≤ 150 chars<br/>& no promo text?}
V7 -->|Fails| FAIL_TITLE[BLOCK: Title validation<br/>failure -- log reason]
V7 -->|Passes| V8{8. Description ≤ 5K<br/>chars & no HTML?}
V8 -->|Fails| FAIL_DESC[BLOCK: Description<br/>validation failure]
V8 -->|Passes| V9{9. Brand field<br/>populated & valid?}
V9 -->|Empty or equals store name| FAIL_BRAND[BLOCK: Invalid brand<br/>-- must be manufacturer brand]
V9 -->|Valid| V10{10. Availability<br/>fresh < 1 hour?}
V10 -->|Stale| REFRESH[Refresh inventory<br/>count from Module 4]
REFRESH --> RECOMPUTE[Recompute availability]
RECOMPUTE --> PASS
V10 -->|Fresh| PASS[ALL CHECKS PASSED]
PASS --> SUBMIT[Submit to Google<br/>Merchant API]
SUBMIT -->|200 OK| SUCCESS[Log success<br/>Update sync timestamp]
SUBMIT -->|Error| RETRY[Retry Pipeline<br/>Section 6.2.3]
FAIL_REQ --> LOG[Log validation failure<br/>to google_sync_log]
FAIL_IMG --> LOG
FAIL_PRICE --> LOG
FAIL_ID --> LOG
FAIL_GTIN --> LOG
FAIL_CAT --> LOG
FAIL_URL --> LOG
FAIL_TITLE --> LOG
FAIL_DESC --> LOG
FAIL_BRAND --> LOG
6.5.9 Google Business Profile Integration
Google Business Profile (GBP) integration is required for Local Inventory Ads. The Merchant Center account must be linked to verified GBP listings so that Google can associate product availability with physical store locations.
GBP-to-Merchant Center Linkage
- Verify store ownership in Google Business Profile for each physical location.
- Link GBP to Merchant Center via the Merchant Center settings panel.
- Map POS locations to GBP listings using the
google_store_codeassigned by GBP. - Enrol locations in Local Inventory Ads via the Merchant Center LIA program.
- Sync store hours from POS location settings to GBP to ensure accuracy.
Data Model: google_business_profile_locations
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants.id |
pos_location_id | UUID | Yes | FK to locations.id (Module 5) |
gbp_place_id | VARCHAR(100) | Yes | Google Places API place ID (e.g., ChIJ...) |
google_store_code | VARCHAR(50) | Yes | Store code used in Local Inventory feed (must match Merchant Center) |
store_name | VARCHAR(200) | Yes | Business name as it appears in GBP |
address_line1 | VARCHAR(255) | Yes | Street address |
address_line2 | VARCHAR(255) | No | Suite, unit, floor |
city | VARCHAR(100) | Yes | City name |
state_province | VARCHAR(100) | Yes | State or province |
postal_code | VARCHAR(20) | Yes | ZIP or postal code |
country_code | CHAR(2) | Yes | ISO 3166-1 alpha-2 (e.g., US, CA) |
phone | VARCHAR(20) | Yes | Primary phone number in E.164 format |
hours_json | JSONB | Yes | Structured store hours (see format below) |
is_verified | BOOLEAN | Yes | Whether GBP ownership is verified – default false |
is_enrolled_lia | BOOLEAN | Yes | Whether Local Inventory Ads are enabled – default false |
last_sync_at | TIMESTAMPTZ | No | Timestamp of most recent GBP sync |
created_at | TIMESTAMPTZ | Yes | Row creation timestamp |
updated_at | TIMESTAMPTZ | Yes | Last modification timestamp |
Store Hours JSON Format
# Example hours_json structure
hours:
monday: { open: "09:00", close: "21:00" }
tuesday: { open: "09:00", close: "21:00" }
wednesday: { open: "09:00", close: "21:00" }
thursday: { open: "09:00", close: "21:00" }
friday: { open: "09:00", close: "22:00" }
saturday: { open: "10:00", close: "22:00" }
sunday: { open: "11:00", close: "18:00" }
special_hours:
- date: "2026-12-25"
closed: true
- date: "2026-12-24"
open: "09:00"
close: "15:00"
Cross-Reference: See Module 5, Section 5.3 for the location management screens where POS locations are configured and mapped to Google Business Profile listings.
6.5.10 Content API Migration Plan
The Content API for Shopping (/content/v2.1/) is being fully replaced by the Merchant API. All POS integrations must target the Merchant API exclusively. This section documents the migration path for any legacy Content API usage.
Migration Timeline
| Step | Action | Deadline | Status |
|---|---|---|---|
| 1 | Audit current Content API usage across all endpoints | January 2026 | Required |
| 2 | Map Content API calls to Merchant API v1 equivalents | February 2026 | Required |
| 3 | Update authentication to use Merchant API token endpoints | March 2026 | Required |
| 4 | Update product data submission to ProductInput resource | April 2026 | Required |
| 5 | Update local inventory to new localInventories endpoints | May 2026 | Required |
| 6 | Full regression testing (product sync, local inventory, notifications) | June 2026 | Required |
| 7 | Production cutover: switch all tenants to Merchant API | July 2026 | Required |
| 8 | Content API fully deprecated – all endpoints cease | August 18, 2026 | Hard deadline |
Key API Mapping: Content API to Merchant API
| Content API (v2.1) | Merchant API (v1) | Breaking Changes |
|---|---|---|
products.insert | productInputs:insert | Write resource renamed to ProductInput; read resource is now Product |
products.get | products.get | Response schema changed; now returns Google-processed version only |
products.list | products.list | Pagination uses pageToken instead of startToken |
products.delete | productInputs:delete | Must target ProductInput resource, not Product |
products.custombatch | Individual calls or batch API | custombatch removed; use standard batch request pattern |
localinventory.insert | localInventories:insert | New URL structure under products/{product}/localInventories |
pos.inventory (legacy) | localInventories:insert | POS-specific endpoint removed; use standard local inventory |
productstatuses.get | productStatuses.get | Response schema updated; new status categories |
Risk Mitigation
| Risk | Probability | Impact | Mitigation |
|---|---|---|---|
| Migration not completed before August 18, 2026 | Medium | Critical – all Google Shopping listings go dark | Start migration in Q1 2026; track weekly in sprint reviews |
| Breaking changes in Merchant API v1 before GA | Low | Medium – requires rework of integration code | Pin to v1beta; monitor Google Merchant API changelog weekly |
| Tenant data inconsistency during cutover | Medium | Medium – temporary product disapprovals | Run Content API and Merchant API in parallel for 2 weeks before final cutover |
| Rate limit changes in new API | Low | Low – may require batching strategy adjustment | Monitor rate limit headers during testing; update google_merchant_sync config as needed |
6.5.11 Reports: Google Merchant Integration
The POS provides five standard reports for monitoring Google Merchant integration health, product approval status, and shopping performance.
Report Catalogue
| Report | Purpose | Key Data Fields | Refresh Frequency |
|---|---|---|---|
| Google Product Status | Track approved, pending, and disapproved products across the feed | sku, google_product_id, status (approved/pending/disapproved), disapproval_reasons[], last_updated | Every 6 hours |
| Google Local Inventory | Monitor store-level availability accuracy and sync latency | store_code, store_name, products_synced, avg_sync_latency_min, stale_count (not synced in 24h), accuracy_pct | Every 4 hours |
| Google Shopping Performance | Surface click, impression, and CTR data from Performance Max campaigns (if available) | product_id, sku, impressions, clicks, ctr_pct, avg_cpc, conversions, revenue | Daily (data from Google Ads API) |
| Google Disapproval Tracker | Track disapproved products with reasons, remediation status, and re-approval dates | sku, product_name, disapproval_reason, disapproved_at, remediation_action, remediation_status (open/in_progress/resolved), re_approved_at | Real-time (via push notifications, Section 6.5.4) |
| Google Feed Health | Overall feed quality score with missing-field analysis and image quality audit | total_products, products_with_all_fields_pct, missing_gtin_count, missing_description_count, low_quality_image_count, feed_health_score (0-100), recommendations[] | Daily |
Feed Health Score Calculation
The Feed Health Score is a composite metric (0–100) calculated as follows:
| Component | Weight | Scoring |
|---|---|---|
| Required fields completeness | 40% | 100 if all products have all required fields; deduct 1 point per product with missing fields (min 0) |
| Image quality compliance | 25% | 100 if all images pass validation; deduct 2 points per product with image issues (min 0) |
| GTIN coverage | 15% | 100 if all products with UPCs have valid GTINs; deduct 1 point per missing GTIN (min 0) |
| Price accuracy | 10% | 100 if all feed prices match website prices; 0 if any mismatch detected |
| Disapproval rate | 10% | 100 if 0% disapproved; 0 if > 10% disapproved; linear between |
feed_health_score:
components:
required_fields:
weight: 0.40
deduction_per_violation: 1
image_quality:
weight: 0.25
deduction_per_violation: 2
gtin_coverage:
weight: 0.15
deduction_per_violation: 1
price_accuracy:
weight: 0.10
scoring: binary # 100 or 0
disapproval_rate:
weight: 0.10
max_acceptable_rate: 0.10
thresholds:
excellent: 90
good: 75
needs_attention: 50
critical: 0
Cross-Reference: See Module 5, Section 5.18 for dashboard widget configuration that surfaces the Feed Health Score on the admin portal home screen.
6.6 Cross-Platform Product Data Requirements
Scope: Ensuring that all product data managed in the POS system meets the validation requirements of every connected external platform simultaneously. Rather than validating per-platform at sync time, the POS enforces a unified “strictest-rule-wins” validation policy at data entry. This guarantees that any product passing POS validation is immediately eligible for listing on Shopify, Amazon, and Google Merchant Center without remediation.
Cross-Reference: See Module 3, Section 3.6 for multi-channel management and channel visibility. See Module 3, Section 3.7 for Shopify-specific sync and field ownership. See Module 5, Section 5.16 for Integration Hub configuration and health monitoring.
Design Principle: The POS system acts as the single source of truth for product data. By enforcing the strictest requirement from any connected platform at the point of data entry, we eliminate the common pattern of “create now, fix later” that leads to suppressed listings, disapproved products, and lost revenue.
6.6.1 Unified Product Data Validation Matrix
The following matrix compares field-level requirements across all three platforms and documents the POS-enforced rule derived from the strictest constraint.
| POS Field | Shopify Requirement | Amazon Requirement | Google Requirement | POS Enforced Rule (Strictest) |
|---|---|---|---|---|
name (title) | Max 255 chars | Max 500 chars | Max 150 chars | Max 150 chars (Google strictest) |
long_description | HTML allowed, no max | HTML allowed, max 2,000 chars (category-dependent) | Max 5,000 chars, plain text preferred | Max 5,000 chars (Google cap; HTML sanitized for Google feed) |
short_description | N/A (uses body) | Max 1,000 chars (bullet points) | N/A | Max 1,000 chars (Amazon bullet point limit) |
primary_image | Any format, 2048x2048 recommended | Min 1000x1000px for zoom eligibility | Min 250x250px (apparel), no watermarks | Min 1000x1000px, no watermarks (Amazon + Google combined) |
base_price | Required | Required per marketplace | Must match landing page price | Required, must match across all channels |
compare_at_price | Optional (strikethrough) | List price (optional) | Optional (sale_price / sale_price_effective_date) | Optional, must be > base_price if set |
barcode (UPC/EAN) | Optional | Required for most categories | Required (GTIN) if exists for product | Required (treat as mandatory for channel eligibility) |
brand | Optional (vendor field) | Required | Required (most categories) | Required |
weight | Optional | Required for FBA fulfillment | Optional (but needed for shipping) | Required (needed for FBA and shipping calculations) |
weight_unit | g, kg, lb, oz | Pounds or kilograms | g, kg, lb, oz | Required, stored in grams, converted per platform |
condition | N/A | Required (New, Refurbished, Used) | Required (new, refurbished, used) | Required (default: new) |
product_type | Optional (free-text) | Required (Amazon Browse Node ID) | Optional (Google Product Category ID) | Required (must map to both Amazon and Google taxonomies) |
sku | Required, unique per store | Required (seller_sku), unique | Required (offerId), max 50 chars | Required, unique per tenant, max 50 chars (Google strictest) |
manufacturer_part_number | N/A | Optional (MPN) | Required if no GTIN assigned | Required if no barcode/GTIN |
color | Optional (variant option) | Required for apparel | Required for apparel | Required for apparel categories |
size | Optional (variant option) | Required for apparel | Required for apparel | Required for apparel categories |
gender | N/A | Required for apparel | Required for apparel | Required for apparel categories |
age_group | N/A | Required for apparel | Required for apparel | Required for apparel categories (adult, kids, toddler, infant, newborn) |
material | N/A | Optional | Recommended for apparel | Recommended (improves listing quality) |
country_of_origin | N/A | Required for some categories | N/A | Required (needed for customs and Amazon compliance) |
Business Rules:
- All text fields are trimmed of leading/trailing whitespace before validation.
- Title (
name) must not contain promotional text (e.g., “FREE SHIPPING”, “SALE”, “BUY NOW”) per Amazon and Google policies. - Price must be > $0.00 for all channel-listed products. Zero-price items are blocked from channel sync.
- If a product fails unified validation, it can still be used for in-store POS sales but is blocked from external channel sync.
6.6.2 Image Requirements Matrix
Product images are the most common reason for listing suppression across platforms. The POS enforces a unified image standard that satisfies all platforms simultaneously.
| Requirement | Shopify | Amazon | POS Enforced Rule | |
|---|---|---|---|---|
| Min resolution | 2048x2048 recommended | 1000x1000 min (zoom eligible) | 250x250 (apparel) / 100x100 (other) | 1000x1000 minimum |
| Max resolution | 4472x4472 | 10000x10000 | N/A | 10000x10000 maximum |
| Max file size | 20MB | 10MB | 16MB | 10MB maximum (Amazon strictest) |
| Formats | JPEG, PNG, GIF | JPEG, PNG, TIFF, GIF | JPEG, PNG, GIF, WebP, BMP, TIFF | JPEG or PNG (universally supported) |
| Color space | sRGB | sRGB | sRGB | sRGB required |
| Background | Any | Pure white (RGB 255,255,255) required for main | White/transparent preferred | White background required (for main image) |
| Watermarks | Allowed | PROHIBITED | PROHIBITED | PROHIBITED |
| Text overlay | Allowed | PROHIBITED on main image | PROHIBITED | PROHIBITED on main image |
| Borders/frames | Allowed | PROHIBITED | PROHIBITED | PROHIBITED |
| Product coverage | N/A | 85% of frame recommended | 75-90% of frame | 85% of frame (Amazon guideline) |
| Main image count | 1 required | 1 required (up to 9 total) | 1 required (up to 11 additional) | 1 required, up to 9 recommended |
| Aspect ratio | Any | 1:1 preferred | Any (1:1 preferred) | 1:1 recommended (square) |
Image Upload Workflow:
flowchart TD
A[Staff Uploads Image] --> B{File Format Check}
B -->|Not JPEG/PNG| C[REJECT: Convert to JPEG or PNG]
B -->|JPEG or PNG| D{Resolution Check}
D -->|< 1000x1000| E[REJECT: Minimum 1000x1000px required]
D -->|>= 1000x1000| F{File Size Check}
F -->|> 10MB| G[WARN: Compress to under 10MB]
F -->|<= 10MB| H{Main Image?}
H -->|Yes| I{Background Check}
H -->|No - Additional| J[PASS: Store Image]
I -->|Not White| K[WARN: White background recommended for main image]
I -->|White| J
G --> L[Auto-Compress & Re-Check]
L --> H
K --> M[Staff Acknowledges Warning]
M --> J
J --> N[Generate Platform Variants]
N --> O[Store Original + Variants]
style C fill:#d32f2f,color:#fff
style E fill:#d32f2f,color:#fff
style G fill:#ff9800,color:#fff
style K fill:#ff9800,color:#fff
style J fill:#2e7d32,color:#fff
Image Storage Strategy:
- Original image stored at upload resolution.
- Platform-optimized variants generated asynchronously: Shopify (2048x2048), Amazon (2000x2000), Google (1200x1200).
- Variant generation uses lossy JPEG compression at quality 85 for file size compliance.
- Images are served via CDN with platform-specific URL patterns.
6.6.3 Pre-Sync Validation Engine
Before pushing any product to any external platform, the validation engine evaluates the product against the unified rules defined above. Validation runs automatically on product save and on-demand before sync operations.
Validation Result Levels:
| Level | Code | Behavior | Description |
|---|---|---|---|
| PASS | PASS | Sync allowed | All required fields present and valid for this platform |
| WARN | WARN | Sync allowed with advisory | Non-blocking issues detected (e.g., recommended fields missing, suboptimal image) |
| FAIL | FAIL | Sync blocked | Required fields missing or invalid; product cannot be listed on this platform |
Product Sync Validation Record:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | FK to products table |
variant_id | UUID | No | FK to product_variants table (NULL for parent-level validation) |
tenant_id | UUID | Yes | FK to tenants table |
platform | Enum | Yes | SHOPIFY, AMAZON, GOOGLE_MERCHANT |
validation_status | Enum | Yes | PASS, WARN, FAIL |
validation_errors | JSON | No | Array of {field, rule, message, severity} objects |
validation_warnings | JSON | No | Array of {field, rule, message, recommendation} objects |
last_validated_at | DateTime | Yes | Timestamp of last validation run |
last_synced_at | DateTime | No | Timestamp of last successful sync to this platform |
sync_status | Enum | Yes | PENDING, SYNCED, FAILED, BLOCKED, NOT_CONFIGURED |
sync_error_message | String(500) | No | Error message from last failed sync attempt |
external_id | String(100) | No | ID on external platform (Shopify product_id, Amazon ASIN, Google offerId) |
external_status | String(50) | No | Status on external platform (active, suppressed, disapproved, pending) |
external_status_reason | Text | No | Platform-reported reason for non-active status |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Validation Error Object Schema:
validation_error:
field: "barcode"
rule: "REQUIRED_FOR_AMAZON"
message: "Barcode (UPC/EAN) is required for Amazon listings in this category"
severity: "FAIL"
platform: "AMAZON"
remediation: "Add a valid UPC or EAN barcode in the product details"
Validation Engine Flow:
flowchart TD
A[Product Change Detected] --> B[Load Product Data]
B --> C[Load Platform Configurations]
C --> D{For Each Enabled Platform}
D --> E[Check Required Fields]
E --> F[Check Field Length Constraints]
F --> G[Check Image Rules]
G --> H[Check Price Consistency]
H --> I[Check Category Mapping]
I --> J[Check Platform-Specific Attributes]
J --> K[Generate Validation Report]
K --> L{Any FAIL Results?}
L -->|Yes| M[Set sync_status = BLOCKED]
L -->|No| N{Any WARN Results?}
N -->|Yes| O[Set validation_status = WARN\nsync_status = PENDING]
N -->|No| P[Set validation_status = PASS\nsync_status = PENDING]
M --> Q[Store Validation Record]
O --> Q
P --> Q
Q --> R[Notify Admin Dashboard]
R --> S{Auto-Sync Enabled?}
S -->|Yes, status = PASS/WARN| T[Queue for Platform Sync]
S -->|No or BLOCKED| U[Await Manual Action]
style M fill:#d32f2f,color:#fff
style O fill:#ff9800,color:#fff
style P fill:#2e7d32,color:#fff
Validation Triggers:
- Product created or updated (automatic)
- Image added, replaced, or removed (automatic)
- Price changed (automatic)
- Manual “Validate All” button in Admin Portal (on-demand)
- Bulk validation via scheduled job (nightly at 2:00 AM tenant time)
Admin Dashboard - Validation Summary View:
| Metric | Description |
|---|---|
| Total Products | Count of all active products in tenant catalog |
| Shopify Ready | Count where Shopify validation = PASS or WARN |
| Amazon Ready | Count where Amazon validation = PASS or WARN |
| Google Ready | Count where Google Merchant validation = PASS or WARN |
| Blocked (per platform) | Count where validation = FAIL, with top 5 failure reasons |
| Needs Attention | Products with WARN status grouped by warning type |
6.6.4 Platform-Specific Product Attributes
Beyond the unified fields, each platform requires or supports additional attributes that are stored in platform-specific extension fields on the product record.
Amazon-Specific Attributes:
| Attribute | Description | POS Storage | Validation |
|---|---|---|---|
product_type | Amazon Browse Node taxonomy classification | amazon_product_type (String) | Must map to valid Amazon category node ID |
bullet_points | Up to 5 key feature bullet points | amazon_bullet_points (JSON array) | Max 5 entries, max 1,000 chars each |
search_terms | Backend search keywords (not visible to customers) | amazon_search_terms (String) | Max 250 bytes total, no ASINs or brand names |
a_plus_content | Enhanced brand content eligibility | amazon_a_plus_eligible (Boolean) | Brand registered sellers only |
fulfillment_channel | Fulfillment method | amazon_fulfillment (Enum) | FBA, FBM, or BOTH |
item_condition_note | Condition details for non-new items | amazon_condition_note (Text) | Required if condition != new, max 1,000 chars |
max_handling_time | Days to ship after order | amazon_handling_days (Integer) | 1-30 days; required for FBM |
restock_date | Expected restock date if out of stock | amazon_restock_date (Date) | Optional; future date only |
Amazon Attribute Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | FK to products table |
tenant_id | UUID | Yes | FK to tenants table |
amazon_product_type | String(100) | Yes | Amazon Browse Node category |
amazon_bullet_points | JSON | No | Array of up to 5 strings, 1000 chars each |
amazon_search_terms | String(250) | No | Backend keywords, space-separated |
amazon_a_plus_eligible | Boolean | Yes | Default: false |
amazon_fulfillment | Enum | Yes | FBA, FBM, BOTH |
amazon_condition_note | Text | No | Required if condition is not new |
amazon_handling_days | Integer | No | Max handling time for FBM orders |
amazon_restock_date | Date | No | Expected restock date |
amazon_asin | String(10) | No | Amazon Standard Identification Number (assigned by Amazon) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Google-Specific Attributes:
| Attribute | Description | POS Storage | Validation |
|---|---|---|---|
google_product_category | Google taxonomy path (numeric ID) | google_category_id (Integer) | Must map to valid Google Product Category |
mpn | Manufacturer part number | manufacturer_part_number (String) | Required if no GTIN; max 70 chars |
additional_image_link | Up to 10 additional images | product_images array (positions 2-11) | Same quality rules as main image |
product_highlight | Up to 10 key feature bullet points | google_highlights (JSON array) | Max 150 chars each, max 10 entries |
local_inventory_attrs | Store-level availability for Local Inventory Ads | Computed from POS inventory per location | Per-location mapping via storeCode |
pickup_method | How customer picks up in-store | google_pickup_method (Enum) | buy, reserve, ship_to_store, not_supported |
pickup_sla | Pickup time estimate | google_pickup_sla (Enum) | same_day, next_day, 2-day, 3-day, 4-day, 5-day, 6-day, multi-week |
custom_label_0 through custom_label_4 | Custom grouping labels for Shopping campaigns | google_custom_labels (JSON) | Max 100 chars each, up to 5 labels |
ads_redirect | Tracking URL for Google Ads | google_ads_redirect (String) | Valid URL, max 2,000 chars |
Google Attribute Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | FK to products table |
tenant_id | UUID | Yes | FK to tenants table |
google_category_id | Integer | Yes | Google Product Category taxonomy ID |
google_category_path | String(500) | No | Human-readable category path (e.g., “Apparel & Accessories > Clothing > Shirts”) |
google_highlights | JSON | No | Array of up to 10 strings, 150 chars each |
google_pickup_method | Enum | No | buy, reserve, ship_to_store, not_supported |
google_pickup_sla | Enum | No | same_day, next_day, 2-day through 6-day, multi-week |
google_custom_labels | JSON | No | Object with keys label_0 through label_4, max 100 chars each |
google_ads_redirect | String(2000) | No | Tracking URL for Google Ads campaigns |
google_offer_id | String(50) | No | Google Merchant Center offer ID (auto-generated from SKU if blank) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Shopify-Specific Attributes:
These fields are owned by Shopify in bidirectional sync mode (see Module 3, Section 3.7). The POS stores them as read-only references.
| Attribute | Description | POS Storage | Ownership |
|---|---|---|---|
seo_title | Meta title for search engine results | shopify_meta_title (String, max 70 chars) | Shopify-Owned |
seo_description | Meta description for search engine results | shopify_meta_description (String, max 320 chars) | Shopify-Owned |
url_handle | URL slug for the product page | shopify_handle (String, max 255 chars) | Shopify-Owned |
metafields | Custom structured data (JSON metafields) | shopify_metafields (JSON) | Shopify-Owned |
collections | Product collection memberships | shopify_collections (JSON array of IDs) | Shopify-Owned |
sales_channels | Channel publishing scope | shopify_channels (JSON array) | Shopify-Owned |
tags | Comma-separated product tags | shopify_tags (Text) | Configurable (POS or Shopify) |
template_suffix | Product page template override | shopify_template_suffix (String) | Shopify-Owned |
Shopify Attribute Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
product_id | UUID | Yes | FK to products table |
tenant_id | UUID | Yes | FK to tenants table |
shopify_product_id | BigInt | No | Shopify internal product ID (assigned after first sync) |
shopify_meta_title | String(70) | No | SEO title |
shopify_meta_description | String(320) | No | SEO description |
shopify_handle | String(255) | No | URL handle/slug |
shopify_metafields | JSON | No | Custom metafield key-value pairs |
shopify_collections | JSON | No | Array of Shopify collection IDs |
shopify_channels | JSON | No | Array of Shopify sales channel names |
shopify_tags | Text | No | Comma-separated tags |
shopify_template_suffix | String(100) | No | Theme template override |
shopify_published_at | DateTime | No | When product was published on Shopify |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Cross-Reference: See Module 3, Section 3.7.2 for field-level ownership model defining which system (POS or Shopify) is authoritative for each field in bidirectional sync mode.
6.7 Cross-Platform Inventory Sync Rules
Scope: Defining the rules, architecture, and failure handling for real-time inventory synchronization between the POS system (source of truth) and all connected external sales channels. This section covers sync latency targets, safety buffer configuration, oversell prevention, channel-specific inventory rules, and failure recovery procedures.
Cross-Reference: See Module 4, Section 4.1 for the POS inventory status model. See Module 4, Section 4.14 for Shopify-specific inventory sync. See Module 3, Section 3.6.3 for channel inventory allocation modes. See Module 5, Section 5.16 for Integration Hub health monitoring.
Design Principle: The POS system maintains a single, authoritative inventory count per product per location. All external channels receive computed available quantities derived from the POS count minus safety buffers and reservations. No external channel can directly modify POS inventory – all inbound changes (e.g., Shopify admin adjustments) are processed through the sync engine with conflict resolution.
6.7.1 Real-Time Inventory Sync Architecture
flowchart TD
POS["POS System\n(Source of Truth)\nSingle Inventory Record\nPer Product Per Location"]
INV_ENGINE["Inventory Sync Engine\n(Event-Driven)"]
SHOP["Shopify\nTarget: < 5s latency\nBidirectional Webhooks"]
AMZ_FBM["Amazon FBM\nTarget: < 2min latency\nAPI Push + SQS Pull"]
AMZ_FBA["Amazon FBA\nRead-Only Monitoring\nAmazon Manages Stock"]
GOOG["Google Merchant Center\nTarget: < 30min processing\nAPI Push (Content API)"]
POS -->|Inventory Event| INV_ENGINE
INV_ENGINE -->|Webhook Push| SHOP
INV_ENGINE -->|Feeds API Push| AMZ_FBM
INV_ENGINE -->|Content API Push| GOOG
INV_ENGINE -.->|Read via SP-API| AMZ_FBA
SHOP -->|Webhook: inventory_levels/update| INV_ENGINE
AMZ_FBM -->|SQS: ANY_OFFER_CHANGED notification| INV_ENGINE
INV_ENGINE -->|Reconciliation| POS
style POS fill:#1565c0,stroke:#0d47a1,color:#fff
style INV_ENGINE fill:#7b2d8e,stroke:#5a1d6e,color:#fff
style SHOP fill:#2e7d32,stroke:#1b5e20,color:#fff
style AMZ_FBM fill:#e65100,stroke:#bf360c,color:#fff
style AMZ_FBA fill:#6c757d,stroke:#495057,color:#fff
style GOOG fill:#c62828,stroke:#b71c1c,color:#fff
Inventory Events That Trigger Sync:
| Event | Source | Channels Notified |
|---|---|---|
| POS Sale completed | POS Terminal | All enabled channels |
| POS Return processed | POS Terminal | All enabled channels |
| Inventory adjustment | Admin Portal | All enabled channels |
| Inventory count reconciliation | Admin Portal | All enabled channels |
| Inter-store transfer (shipped) | Source location | All enabled channels (source location) |
| Inter-store transfer (received) | Destination location | All enabled channels (destination location) |
| Purchase order received | Receiving module | All enabled channels (receiving location) |
| Online order reserved | Shopify/Amazon | Remaining channels |
| Online order cancelled | Shopify/Amazon | All enabled channels (restore qty) |
Sync Latency Targets:
| Channel | Sync Method | Target Latency | Reconciliation Frequency | Max Acceptable Lag |
|---|---|---|---|---|
| Shopify | Webhooks (bidirectional) | < 5 seconds | Every 15 minutes | 60 seconds |
| Amazon FBM | SP-API push + SQS pull | < 2 minutes | Every 30 minutes | 10 minutes |
| Amazon FBA | Read-only monitoring (SP-API) | N/A (Amazon manages) | Every 4 hours | N/A |
| Google Merchant | Content API push | < 30 minutes (Google processing time) | Every 6 hours | 60 minutes |
Reconciliation Process:
- At each reconciliation interval, the sync engine compares POS quantities against platform-reported quantities.
- Discrepancies are logged in the Integration Sync Log (Module 5, Section 5.16.4).
- If discrepancy exceeds the configured threshold (default: 5 units), an admin alert is triggered.
- Auto-correction pushes POS quantity to the platform (POS always wins in reconciliation).
6.7.2 Safety Buffer Configuration
Safety buffers prevent overselling by withholding a configurable number of units from external channel listings. This provides a cushion for in-store sales, processing delays, and inventory inaccuracies.
Primary Formula:
Channel Available Qty = POS Available Qty - Safety Buffer
If Channel Available Qty < min_channel_qty → Show as out_of_stock on that channel
If max_channel_qty is set → Channel Available Qty = MIN(Channel Available Qty, max_channel_qty)
Safety Buffer Settings:
| Setting | Description | Default | Per-Product Override | Per-Channel Override |
|---|---|---|---|---|
safety_buffer_qty | Fixed units withheld from channel listing | 0 | Yes | Yes |
safety_buffer_pct | Percentage of POS qty withheld (alternative to fixed) | NULL | Yes | Yes |
buffer_calculation | How the buffer is computed | FIXED | Yes | Yes |
channel_warehouse_id | Specific POS location(s) feeding this channel | NULL (all locations) | No | Yes |
min_channel_qty | Minimum qty to display on channel; below this = out_of_stock | 1 | Yes | Yes |
max_channel_qty | Maximum qty shown on channel (cap) | NULL (no cap) | Yes | Yes |
Buffer Calculation Modes:
| Mode | Formula | Use Case | Example (POS Qty = 20, Buffer = 3) |
|---|---|---|---|
FIXED | Channel Qty = POS Qty - buffer_qty | Simple fixed reserve for walk-in customers | 20 - 3 = 17 listed |
PERCENTAGE | Channel Qty = POS Qty - CEIL(POS Qty * buffer_pct / 100) | Proportional reserve that scales with stock level | 20 - CEIL(20 * 15%) = 20 - 3 = 17 listed |
MIN_RESERVE | Channel Qty = MAX(0, POS Qty - buffer_qty) | Floor-based reserve (never goes negative) | 20 - 3 = 17 listed (or 0 if POS Qty < buffer) |
Buffer Priority Resolution:
- Product-specific + Channel-specific override (highest priority)
- Product-specific override (applies to all channels)
- Channel-specific default (applies to all products on that channel)
- Tenant-wide default (lowest priority)
Safety Buffer Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
product_id | UUID | No | FK to products table; NULL = tenant-wide or channel-wide default |
variant_id | UUID | No | FK to product_variants; NULL = applies to all variants of product |
channel | Enum | Yes | SHOPIFY, AMAZON_FBM, GOOGLE_MERCHANT, ALL |
safety_buffer_qty | Integer | Yes | Fixed units to withhold (default: 0) |
safety_buffer_pct | Decimal(5,2) | No | Percentage buffer (alternative to fixed); NULL if using fixed |
buffer_calculation | Enum | Yes | FIXED, PERCENTAGE, MIN_RESERVE |
min_channel_qty | Integer | Yes | Below this threshold = show as out_of_stock (default: 1) |
max_channel_qty | Integer | No | Cap on listed quantity; NULL = unlimited |
channel_warehouse_id | UUID | No | FK to locations; specific location for this channel; NULL = aggregate all locations |
is_active | Boolean | Yes | Whether this buffer rule is active (default: true) |
priority | Integer | Yes | Resolution priority (lower = higher priority); auto-calculated |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Recommended Buffer Defaults by Channel:
| Channel | Recommended Buffer | Rationale |
|---|---|---|
| Shopify | 0-2 units (fixed) | Low latency (< 5s) reduces oversell risk |
| Amazon FBM | 5-10% (percentage) | 2-minute sync lag + high velocity = higher risk |
| Google Merchant | 10-15% (percentage) | 30-minute processing delay = highest risk |
6.7.3 Oversell Prevention Rules
Overselling occurs when two or more channels sell the last available units simultaneously before inventory sync propagates the decrement. The POS system uses a reserve-on-order model with first-commit-wins conflict resolution.
Core Principles:
- All channels sync from a single POS inventory source (no shadow inventory).
- When an order is received from ANY channel, the POS immediately creates a reservation (soft lock) against the inventory.
- If two channels attempt to reserve the last unit simultaneously, the first transaction to commit wins; the second receives an insufficient stock response.
- Safety buffers provide a cushion during sync propagation windows.
- During offline mode, channels receive the last-known inventory quantity; the safety buffer provides protection against stale data.
Oversell Prevention Sequence:
sequenceDiagram
autonumber
participant SHOP as Shopify Store
participant POS as POS Inventory Engine
participant AMZ as Amazon Marketplace
participant DB as Database
Note over SHOP,AMZ: Scenario: 1 unit available, 2 simultaneous orders
SHOP->>POS: Order webhook (product X, qty: 1)
AMZ->>POS: Order notification (product X, qty: 1)
POS->>DB: BEGIN TRANSACTION (Shopify order)
DB-->>POS: Lock acquired on inventory row
POS->>DB: Reserve 1 unit (qty_available: 1 → 0, qty_reserved: +1)
DB-->>POS: COMMIT SUCCESS
POS->>DB: BEGIN TRANSACTION (Amazon order)
DB-->>POS: Lock acquired on inventory row
POS->>DB: Attempt reserve 1 unit (qty_available: 0)
DB-->>POS: INSUFFICIENT STOCK - ROLLBACK
POS-->>SHOP: Order confirmed - fulfillment pending
POS-->>AMZ: Reject order - insufficient stock
POS->>SHOP: Inventory update: qty = 0
POS->>AMZ: Inventory update: qty = 0
Note over SHOP,AMZ: Amazon order auto-cancelled or backordered per tenant config
Conflict Resolution Policies:
| Scenario | Resolution | Tenant Configurable |
|---|---|---|
| Two channels sell last unit simultaneously | First-commit wins (database row lock) | No (system behavior) |
| Losing channel order | Auto-cancel with customer notification OR backorder | Yes (per channel) |
| POS sale conflicts with online order | POS sale always wins (staff has physical product) | No (system behavior) |
| Inventory goes negative (edge case) | Alert admin, freeze channel sync, require manual resolution | No (safety mechanism) |
| Stale inventory during offline mode | Safety buffer absorbs; reconcile on reconnection | Yes (buffer size) |
Backorder Policy Options (per channel, per tenant):
| Policy | Behavior | Use Case |
|---|---|---|
AUTO_CANCEL | Automatically cancel the losing order and notify customer | Default for most retailers |
BACKORDER | Accept the order as backorder; fulfill when stock arrives | High-value or made-to-order products |
MANUAL_REVIEW | Hold the order for staff decision | Conservative approach |
6.7.4 Channel-Specific Inventory Rules
Each external channel has unique inventory sync behaviors, API constraints, and recommended configurations.
Amazon Inventory Rules:
| Rule | Value | Notes |
|---|---|---|
| FBA inventory tracking | Separate – Amazon manages physical stock | POS monitors FBA levels via SP-API getInventorySummaries; does not push to FBA |
| FBM inventory sync | From POS locations via Feeds API | Uses JSON_LISTINGS_FEED for inventory updates |
| FBM sync trigger | Every POS inventory event | Push via SP-API submitFeed or patchListingsItem |
| Order polling frequency | Every 2 minutes | New orders checked via getOrders SP-API endpoint |
| Safety buffer recommendation | 5-10% (percentage mode) | Higher buffer for slow-sync or high-velocity items |
| Handling time | Configurable per product (default: 2 days) | Affects customer delivery expectation |
| Multi-location support | Channel warehouse mapping | Map specific POS location(s) to Amazon FBM fulfillment |
| Throttling limits | 10 requests/sec for Feeds API | Batch updates to stay within rate limits |
| Quantity cap | Amazon shows “In Stock” for qty > 0 | Exact quantity not displayed to customers on most categories |
Amazon Inventory Sync Flow:
sequenceDiagram
autonumber
participant POS as POS System
participant ENGINE as Sync Engine
participant SP as Amazon SP-API
participant SQS as Amazon SQS
Note over POS,SQS: Outbound: POS → Amazon
POS->>ENGINE: Inventory event (product X, location A, new qty: 15)
ENGINE->>ENGINE: Calculate buffer (15 - 10% = 13)
ENGINE->>SP: patchListingsItem (sku: X, qty: 13)
SP-->>ENGINE: 200 OK
Note over POS,SQS: Inbound: Amazon → POS
SQS->>ENGINE: ANY_OFFER_CHANGED notification
ENGINE->>SP: getListingsItem (sku: X)
SP-->>ENGINE: Listing data with fulfillable_qty
ENGINE->>POS: Reconcile if discrepancy detected
Google Merchant Inventory Rules:
| Rule | Value | Notes |
|---|---|---|
| Sync scope | Local inventory per storeCode | Each POS location maps to a Google storeCode for Local Inventory Ads |
| Online inventory | Aggregated or warehouse-specific | Configurable: aggregate all locations or use designated warehouse |
| Processing delay | Up to 30 minutes after API submission | Google processes updates asynchronously |
| Safety buffer recommendation | 10-15% (percentage mode) | Higher buffer to account for processing delay |
| Availability values | in_stock, out_of_stock, preorder, backorder | Computed from POS qty and buffer rules |
| Update method | Content API localInventory.insert for local; products.insert for online | Different endpoints for local vs. online inventory |
| Update frequency | On every POS inventory event + 2x daily full sync | Event-driven + full reconciliation ensures accuracy |
| Quantity precision | Whole numbers only | Fractional quantities rounded down |
| Sale price sync | sale_price + sale_price_effective_date | Must match actual price on landing page |
Google Merchant Availability Mapping:
| POS Qty (after buffer) | Google Availability | Additional Fields |
|---|---|---|
| qty >= min_channel_qty | in_stock | quantity: actual qty |
| qty = 0 and restock date set | backorder | availability_date: restock date |
| qty = 0 and preorder flag | preorder | availability_date: release date |
| qty = 0 (default) | out_of_stock | – |
Shopify Inventory Rules:
| Rule | Value | Notes |
|---|---|---|
| Sync mode | Real-time bidirectional via webhooks | < 5s target latency |
| Granularity | Per-location | Each POS location maps 1:1 to a Shopify location |
| Location mapping | pos_location_id ↔ shopify_location_id | Configured in Integration Hub (Module 5, Section 5.16.3) |
| Webhook events | inventory_levels/update (inbound), Inventory Level Set API (outbound) | Bidirectional sync |
| Reconciliation | Every 15 minutes | Full inventory comparison per location to catch missed webhooks |
| Safety buffer | Optional (low latency reduces need) | Default: 0 for Shopify channel |
| Track inventory | MUST be enabled for all synced products | Products without inventory tracking are skipped |
| Multi-location | Supported natively | Shopify supports multiple inventory locations |
| Negative inventory | Blocked in POS; Shopify setting must match | Ensure “Allow negative inventory” is disabled in Shopify |
| Inventory policy | deny (stop selling at 0) or continue (oversell allowed) | Synced from POS channel config; deny recommended |
Shopify Location Mapping Data Model:
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
pos_location_id | UUID | Yes | FK to POS locations table |
shopify_location_id | BigInt | Yes | Shopify location ID |
shopify_location_name | String(100) | No | Cached Shopify location name for display |
sync_enabled | Boolean | Yes | Whether inventory syncs for this location pair (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
6.7.5 Sync Failure Handling
Inventory sync failures are critical because stale inventory data directly causes overselling or lost sales (showing out-of-stock when units are available). The system implements a tiered failure response with automatic escalation.
Retry Strategy:
- Attempt 1: Immediate retry after 5 seconds
- Attempt 2: Retry after 15 seconds (3x backoff)
- Attempt 3: Retry after 45 seconds (3x backoff)
- After 3 failures: Message moved to dead letter queue (DLQ)
- DLQ processor retries every 5 minutes for up to 2 hours
- After 2 hours: Admin escalation and channel freeze
Failure Escalation Timeline:
| Failure Duration | Action | Channel Effect | Admin Notification |
|---|---|---|---|
| 0-5 seconds | Automatic retry (attempt 1) | No impact | None |
| 5-15 seconds | Retry with backoff (attempt 2) | No impact | None |
| 15-45 seconds | Final retry (attempt 3) | Minimal delay | None |
| 45s - 5 minutes | Dead letter queue, auto-retry every 5 min | Slight staleness possible | None |
| 5 - 30 minutes | DLQ retries continue | Channel qty may be stale | Warning badge on Integration Hub |
| 30 min - 2 hours | Admin alert (email + in-app notification) | Channel qty is stale | Alert: “Inventory sync failing for [channel]” |
| > 2 hours | Channel freeze: set all products to out_of_stock | Products shown as unavailable | Critical alert: “Channel [X] frozen - manual intervention required” |
Sync Failure Sequence Diagram:
sequenceDiagram
autonumber
participant POS as POS System
participant ENGINE as Sync Engine
participant PLATFORM as External Platform
participant DLQ as Dead Letter Queue
participant ADMIN as Admin Dashboard
POS->>ENGINE: Inventory change event
ENGINE->>PLATFORM: Push inventory update (attempt 1)
PLATFORM-->>ENGINE: ERROR (timeout/5xx)
Note over ENGINE: Wait 5 seconds
ENGINE->>PLATFORM: Retry (attempt 2)
PLATFORM-->>ENGINE: ERROR (timeout/5xx)
Note over ENGINE: Wait 15 seconds
ENGINE->>PLATFORM: Retry (attempt 3)
PLATFORM-->>ENGINE: ERROR (timeout/5xx)
ENGINE->>DLQ: Move to dead letter queue
ENGINE->>ENGINE: Log failure in sync_log
loop Every 5 minutes for up to 2 hours
DLQ->>ENGINE: Dequeue message
ENGINE->>PLATFORM: Retry sync
alt Success
PLATFORM-->>ENGINE: 200 OK
ENGINE->>POS: Update sync_status = SYNCED
else Still failing
PLATFORM-->>ENGINE: ERROR
ENGINE->>DLQ: Re-queue message
end
end
Note over ENGINE: 30 minutes elapsed
ENGINE->>ADMIN: Warning: Sync failing for 30+ minutes
Note over ENGINE: 2 hours elapsed
ENGINE->>ADMIN: CRITICAL: Channel frozen
ENGINE->>PLATFORM: Set all products to out_of_stock
ENGINE->>POS: Update sync_status = FROZEN for all products on channel
Failure Types and Handling:
| Failure Type | Detection | Recovery | Auto-Resolve |
|---|---|---|---|
| Network timeout | HTTP timeout (30s) | Retry with backoff | Yes (transient) |
| Rate limiting (429) | HTTP 429 response | Exponential backoff, respect Retry-After header | Yes |
| Authentication expired | HTTP 401/403 | Refresh OAuth token; if fails, alert admin | Partial (token refresh is automatic) |
| Platform outage | HTTP 5xx repeated | DLQ + escalation timeline | Yes (when platform recovers) |
| Invalid data (400) | HTTP 400 with error details | Log error, skip item, alert admin | No (requires data fix) |
| Webhook delivery failure | No acknowledgment from POS | Platform retries (Shopify: up to 19 times over 48h) | Yes (platform-managed retry) |
Manual Recovery Tools:
| Tool | Location | Function |
|---|---|---|
| Resync Single Product | Product Detail > Integration Tab | Push current POS inventory to all channels for one product |
| Resync All Products | Integration Hub > Channel Card | Full inventory push to one channel |
| Unfreeze Channel | Integration Hub > Channel Card | Remove out_of_stock freeze and resume normal sync |
| View Failed Syncs | Integration Hub > Sync Log | Filter by status = FAILED, view error details, retry individually |
| Force Reconciliation | Integration Hub > Channel Card | Trigger immediate full reconciliation outside normal schedule |
Sync Health Monitoring Dashboard Metrics:
| Metric | Description | Alert Threshold |
|---|---|---|
| Sync success rate (1h) | Percentage of successful syncs in last hour | < 95% = Warning, < 80% = Critical |
| Average sync latency | Mean time from POS event to platform confirmation | > 2x target latency = Warning |
| DLQ depth | Number of messages in dead letter queue | > 50 = Warning, > 200 = Critical |
| Reconciliation discrepancies | Count of qty mismatches found in last reconciliation | > 10 = Warning, > 50 = Critical |
| Time since last successful sync | Elapsed time since last confirmed sync per channel | > 15 min (Shopify), > 30 min (Amazon), > 2h (Google) = Warning |
Business Rules:
- Sync failures for a single product do not affect other products on the same channel.
- Channel freeze (out_of_stock) is a safety mechanism, not a punitive one. It prevents overselling during extended outages.
- Unfreezing a channel triggers an immediate full reconciliation to ensure accuracy before resuming normal sync.
- All sync failures are logged in the Integration Sync Log (Module 5, Section 5.16.4) with full error details for troubleshooting.
- Tenants can configure per-channel freeze thresholds (override the default 2-hour threshold) in the Integration Hub settings.
6.7.6 Reports: Cross-Platform Inventory Sync
| Report | Purpose | Key Data Fields |
|---|---|---|
| Sync Health Summary | Overall sync performance across all channels | Channel, success rate %, avg latency, DLQ depth, last successful sync, status |
| Oversell Incident Report | Track instances where overselling occurred despite safeguards | Date, channel, product, POS qty at time, channel qty at time, order details, root cause |
| Safety Buffer Effectiveness | Analyze whether buffers are appropriately sized | Channel, buffer setting, oversell incidents, missed sales (buffer too high), recommendation |
| Sync Failure Analysis | Detailed breakdown of sync failures by type and channel | Channel, failure type, count, avg resolution time, auto-resolved %, manual intervention count |
| Inventory Discrepancy Log | Reconciliation findings showing POS vs. channel qty differences | Product, channel, POS qty, channel qty, discrepancy, reconciliation action, timestamp |
| Channel Freeze History | Track channel freeze events and their impact | Channel, freeze start, freeze end, duration, products affected, estimated lost revenue |
End of Module 6: Integrations – Sections 6.6 and 6.7
6.8 Payment Processor Integration
Scope: Consolidation of payment integration architecture, terminal communication, processor configuration, failure handling, and batch settlement. This section brings together payment-related content previously distributed across Module 1 (Sections 1.18.1-1.18.3) and Module 5 (Section 5.11.3) into a unified integration specification.
Cross-Reference: The payment flow sequence diagram (card tap/insert, refund via token) remains in Module 1, Section 1.18. Payment method registry (CASH, CREDIT_CARD, GIFT_CARD, etc.) and per-location payment configuration remain in Module 5, Section 5.11.1-5.11.2.
6.8.1 SAQ-A Architecture
The POS system uses a semi-integrated payment architecture that achieves PCI SAQ-A compliance – the simplest and least burdensome PCI self-assessment level. In this architecture, card data never enters or traverses the POS application. The payment terminal communicates directly with the payment processor, and only non-sensitive tokens and metadata flow back to the POS system.
Data the POS System Stores (Safe)
| Field | Type | Example | Purpose |
|---|---|---|---|
transaction_id | UUID | a1b2c3d4-... | Internal payment reference |
payment_token | String(255) | tok_1Nq... | Processor-issued token for refunds and voids |
approval_code | String(20) | AUTH4829 | Authorization code from issuing bank |
masked_card_number | String(10) | ****1234 | Last 4 digits only |
card_brand | Enum | VISA | Visa, Mastercard, Amex, Discover |
entry_method | Enum | TAP | CHIP, TAP, SWIPE, MANUAL |
terminal_id | String(50) | TRM-001-GM | Which physical terminal processed the payment |
timestamp | DateTime | 2026-02-17T14:32:00Z | When the authorization was obtained |
amount | Decimal(10,2) | 45.00 | Authorized or settled amount |
PCI Prohibited Data (NEVER Stored)
| Data Element | PCI Requirement | Risk if Stored |
|---|---|---|
| Full PAN (16-digit card number) | Prohibited under SAQ-A | Immediate PCI non-compliance; liability for fraud |
| CVV / CVC (3-4 digit security code) | Prohibited post-authorization | Stored CVV is grounds for PCI ban by acquirer |
| Track 1 / Track 2 data (magnetic stripe) | Prohibited post-authorization | Enables card cloning |
| PIN block (encrypted PIN) | Prohibited post-authorization | Enables unauthorized transactions |
| Raw EMV data (chip cryptogram) | Prohibited post-authorization | Enables replay attacks |
Architecture Principle: The POS backend sends the payment amount and order reference to the terminal. The terminal handles all card interaction. The processor returns a token. The POS stores only the token. This ensures the POS application, its database, and its network are entirely out of PCI cardholder data scope.
6.8.2 Terminal Communication Protocol
terminal_integration:
# Communication method
protocol: "cloud_api" # Terminal vendor's cloud service
# Timeout settings
payment_timeout_seconds: 60
connection_timeout_seconds: 10
# Terminal state machine
states:
- IDLE: "Ready for transaction"
- WAITING_FOR_CARD: "Display amount, await tap/insert"
- PROCESSING: "Communicating with processor"
- APPROVED: "Transaction successful"
- DECLINED: "Transaction declined"
- ERROR: "Communication or hardware error"
- CANCELLED: "Customer or staff cancelled"
# Error handling
on_timeout: "prompt_retry_or_cancel"
on_decline: "display_reason_allow_retry"
on_error: "log_and_alert_manager"
# Void window (same-day before batch)
same_day_void: true
batch_close_time: "23:00" # Auto-batch at 11 PM
Terminal State Machine
stateDiagram-v2
[*] --> IDLE: Terminal powered on
IDLE --> WAITING_FOR_CARD: Payment request received
WAITING_FOR_CARD --> PROCESSING: Card presented
WAITING_FOR_CARD --> CANCELLED: Staff/customer cancels
WAITING_FOR_CARD --> IDLE: Timeout (60s)
PROCESSING --> APPROVED: Processor approves
PROCESSING --> DECLINED: Processor declines
PROCESSING --> ERROR: Communication failure
APPROVED --> IDLE: Transaction complete
DECLINED --> IDLE: Staff acknowledges
CANCELLED --> IDLE: Reset terminal
ERROR --> IDLE: Staff acknowledges
6.8.3 Processor Configuration (from 5.11.3)
External payment processors handle card transactions and third-party financing. Each processor is configured with encrypted credentials, terminal mappings, and environment settings.
Payment Processor Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
name | String(100) | Yes | Processor display name (e.g., “CardConnect”, “Square”, “Affirm”) |
processor_type | Enum | Yes | CARD, FINANCING |
api_key | String(500) | Yes | Encrypted API key or token (AES-256 encrypted at rest) |
api_secret | String(500) | No | Encrypted API secret (for processors requiring key + secret) |
merchant_id | String(100) | Yes | Merchant account identifier with the processor |
webhook_url | String(500) | No | URL for processor to send asynchronous notifications (refund confirmations, chargebacks) |
test_mode | Boolean | Yes | true for sandbox/test environment, false for production (default: true) |
is_active | Boolean | Yes | Soft-delete flag (default: true) |
config_json | JSON | No | Processor-specific configuration (e.g., Affirm min/max order amounts, supported card brands) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Processor Terminal Mapping
Each payment terminal (physical card reader) is associated with a processor and a location.
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
processor_id | UUID | Yes | FK to payment_processors table |
terminal_id | String(50) | Yes | Terminal serial number or identifier assigned by the processor |
ip_address | String(45) | No | Local IP address of the terminal (for IP-connected terminals) |
port | Integer | No | Port number for terminal communication (default: 9000) |
location_id | UUID | Yes | FK to locations table – physical location of this terminal |
is_active | Boolean | Yes | Soft-delete flag (default: true) |
created_at | DateTime | Yes | Record creation timestamp |
Business Rules:
- A tenant may have multiple processors of the same type (e.g., one CARD processor for in-store terminals and another for e-commerce).
test_modeallows the administrator to toggle between sandbox and production without changing API keys (processors typically issue separate sandbox and production keys).- Credentials (
api_key,api_secret) are encrypted at rest using AES-256 and never returned in API responses. The Admin Portal displays only a masked preview (e.g.,sk_live_****ABCD). - Before activating a processor (
test_mode = false), the system performs a validation handshake with the processor API to confirm credentials are valid.
6.8.4 Failure Handling
sequenceDiagram
autonumber
participant U as Staff
participant UI as POS UI
participant API as Backend
participant TERM as Payment Terminal
participant PROC as Processor
U->>UI: Click "Pay by Card"
UI->>API: POST /payments/initiate
API->>TERM: Send Payment Request
alt Terminal Timeout (60s)
TERM--xAPI: No Response
API-->>UI: "Terminal not responding"
UI-->>U: Options: Retry | Different Terminal | Cash | Cancel
alt Staff selects Retry
U->>UI: Click "Retry"
UI->>API: POST /payments/initiate (same order)
API->>TERM: Send Payment Request (attempt 2)
else Staff selects Different Terminal
U->>UI: Select alternate terminal
UI->>API: POST /payments/initiate (new terminal_id)
API->>TERM: Send to alternate terminal
else Staff selects Cash
U->>UI: Switch to Cash payment
Note right of UI: Proceeds to cash drawer flow
else Staff selects Cancel
U->>UI: Return to cart
end
else Card Declined
TERM->>PROC: Card data (encrypted)
PROC-->>TERM: DECLINED (reason: Insufficient Funds)
TERM-->>API: Decline response
API-->>UI: "Card Declined: Insufficient Funds"
UI-->>U: Options: Try Another Card | Cash | Cancel
API->>API: Log decline (reason, terminal, timestamp)
else Terminal Hardware Error
TERM-->>API: ERROR (Hardware Issue)
API-->>UI: "Terminal Error"
UI-->>U: Options: Different Terminal | Cash | Cancel
API->>API: Log error, create manager alert
Note right of API: Alert sent to on-duty manager
end
Failure Logging:
- All payment failures are recorded in the
payment_attemptslog with:order_id,terminal_id,attempt_number,failure_type(TIMEOUT, DECLINED, ERROR),decline_reason,timestamp. - Consecutive failures on the same terminal (3+ within 15 minutes) trigger a terminal health alert visible on the Integration Health Dashboard.
6.8.5 Batch Settlement
Batch settlement closes the day’s card transactions and initiates fund transfer from the payment processor to the merchant’s bank account.
Settlement Schedule:
- Auto-batch executes at a configurable time (default: 23:00 local time per location timezone).
- Manual batch close is available to managers via the Admin Portal.
- Batch close is per-processor, per-location (each location settles its own terminals).
Settlement Process:
| Step | Action | System Behavior |
|---|---|---|
| 1 | Trigger batch close | System sends batch close command to processor API |
| 2 | Processor responds | Processor returns batch summary (transaction count, total amount) |
| 3 | Reconciliation | POS compares processor batch total against local transaction records |
| 4 | Variance detection | If totals differ by more than $0.01, a reconciliation alert is created |
| 5 | Settlement report | Daily settlement report generated and available in Reports module |
| 6 | Fund transfer | Processor initiates ACH/wire to merchant bank (1-3 business days) |
Reconciliation Rules:
- POS-side total is calculated as: SUM(authorized amounts) - SUM(voided amounts) - SUM(refunded amounts) for the batch period.
- Processor-side total is received via the batch close API response.
- Variance <= $0.01 is auto-accepted (rounding tolerance).
- Variance > $0.01 generates a
RECONCILIATION_VARIANCEalert and requires manager review.
6.8.6 Reports: Payment Integration
| Report | Purpose | Key Data Fields |
|---|---|---|
| Payment Terminal Performance | Monitor terminal health and throughput | Terminal ID, transaction count, avg response time (ms), error rate (%), decline rate (%), uptime (%) |
| Decline Rate Report | Analyze payment failure patterns | Decline reason, frequency, terminal ID, time of day, retry success rate (%), card brand breakdown |
| Batch Settlement Report | Daily batch close summary and reconciliation | Batch date, transaction count, total amount, processor total, variance, settlement status (SETTLED / PENDING / VARIANCE) |
| Chargeback Tracking | Monitor disputes and chargebacks | Chargeback ID, original transaction, amount, reason code, deadline, status (OPEN / WON / LOST) |
6.9 Email Provider Integration
Scope: Consolidation of email provider configuration for all outbound transactional and notification emails. This section brings together provider setup previously in Module 5, Section 5.15.1.
Cross-Reference: Email template registry, template data model, merge field definitions, and per-template enablement controls remain in Module 5, Section 5.15.2-5.15.4. See Module 2, Section 2.9 for customer communication preferences.
6.9.1 Provider Configuration (from 5.15.1)
Each tenant configures exactly one email provider. The provider handles all outbound transactional and notification emails for the tenant. Three provider types are supported.
Email Config Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
provider_type | Enum | Yes | SMTP, SENDGRID, MAILGUN |
smtp_host | String(255) | Conditional | SMTP server hostname (required when provider_type = SMTP) |
smtp_port | Integer | Conditional | SMTP server port: 587 (TLS) or 465 (SSL) (required when provider_type = SMTP) |
smtp_username | String(255) | Conditional | SMTP authentication username (required when provider_type = SMTP) |
smtp_password_encrypted | String(500) | Conditional | AES-256 encrypted SMTP password (required when provider_type = SMTP) |
api_key_encrypted | String(500) | Conditional | AES-256 encrypted API key (required when provider_type = SENDGRID or MAILGUN) |
api_region | Enum | No | US, EU – Mailgun region (default: US) |
sender_email | String(255) | Yes | “From” address for all outbound emails (e.g., noreply@nexusclothing.com) |
sender_name | String(100) | Yes | “From” display name (e.g., “Nexus Clothing”) |
reply_to_email | String(255) | No | “Reply-To” address; defaults to sender_email if NULL |
daily_send_limit | Integer | No | Maximum emails per 24-hour period (0 = unlimited; default: 0) |
is_verified | Boolean | Yes | Whether the configuration has been verified via test email (default: false) |
verified_at | DateTime | No | Timestamp of last successful verification |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Business Rules:
- Only one active email configuration per tenant. Updating the provider replaces the previous configuration; the old record is soft-deleted for audit.
- The
is_verifiedflag is set totrueonly after a successful test email delivery. The Admin Portal displays a warning banner if the configuration is unverified. - All sensitive fields (
smtp_password_encrypted,api_key_encrypted) are encrypted at rest using AES-256. They are never returned in API GET responses – only a masked indicator (e.g.,********) is shown. - When
provider_type = SMTP, the system attempts a TLS handshake during verification. If TLS fails, verification fails with a specific error message.
6.9.2 Delivery Monitoring
The system tracks delivery outcomes for all outbound emails to detect provider issues early and ensure transactional emails reach their intended recipients.
Delivery Status Tracking
| Status | Description | Source |
|---|---|---|
QUEUED | Email accepted by POS, awaiting provider submission | Internal |
SENT | Email accepted by provider for delivery | Provider API response |
DELIVERED | Email delivered to recipient inbox | Provider webhook |
BOUNCED | Email rejected by recipient mail server | Provider webhook |
SOFT_BOUNCE | Temporary delivery failure (mailbox full, server down) | Provider webhook |
SPAM_COMPLAINT | Recipient marked email as spam | Provider webhook |
FAILED | Provider rejected email (invalid sender, quota exceeded) | Provider API response |
Delivery Log Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
template_code | String(50) | Yes | Email template code (e.g., TMPL-REFUND-CONFIRMATION) |
recipient_email | String(255) | Yes | Recipient email address |
status | Enum | Yes | Delivery status from table above |
provider_message_id | String(255) | No | Message ID returned by provider (for tracking) |
error_message | String(500) | No | Error details for failed/bounced emails |
sent_at | DateTime | Yes | Timestamp when email was submitted to provider |
delivered_at | DateTime | No | Timestamp when delivery was confirmed |
created_at | DateTime | Yes | Record creation timestamp |
Provider Webhook Integration:
| Provider | Webhook Mechanism | Endpoint | Events Tracked |
|---|---|---|---|
| SendGrid | Event Webhook | /api/webhooks/sendgrid/events | delivered, bounced, dropped, spam_report |
| Mailgun | Events API / Webhooks | /api/webhooks/mailgun/events | delivered, failed, complained |
| SMTP | None (poll-based) | N/A | Delivery inferred from absence of bounce within 24 hours |
Monitoring Alerts:
- Bounce rate exceeding 5% in a rolling 24-hour window triggers a dashboard alert to tenant administrators.
- Three consecutive delivery failures to the same recipient suppress further emails to that address until an administrator reviews and clears the suppression.
- Daily send limit approaching 80% capacity triggers a warning notification.
6.10 Carrier & Shipping Integration
Scope: Abstract carrier interface, supported carrier configuration, and shipping data model for ship-to-customer fulfillment. This section defines the integration framework for rate lookup, label generation, address validation, and shipment tracking.
Cross-Reference: The ship-to-customer sales workflow, store assignment logic, and pick-pack-ship process remain in Module 1, Section 1.7.3. See Module 4, Section 4.14 for online order fulfillment inventory logic.
6.10.1 Abstract Carrier Interface
All carrier integrations implement a common interface that abstracts provider-specific API differences. This enables the POS system to support multiple carriers without changes to business logic.
classDiagram
class ICarrierProvider {
<<interface>>
+GetRates(origin, destination, package) ShippingRate[]
+CreateLabel(shipment) Label
+GetTracking(trackingNumber) TrackingStatus
+CancelShipment(shipmentId) bool
+ValidateAddress(address) AddressValidation
}
class UPSProvider {
+GetRates()
+CreateLabel()
+GetTracking()
+CancelShipment()
+ValidateAddress()
}
class FedExProvider {
+GetRates()
+CreateLabel()
+GetTracking()
+CancelShipment()
+ValidateAddress()
}
class USPSProvider {
+GetRates()
+CreateLabel()
+GetTracking()
+CancelShipment()
+ValidateAddress()
}
class AmazonShippingProvider {
+GetRates()
+CreateLabel()
+GetTracking()
+CancelShipment()
+ValidateAddress()
}
ICarrierProvider <|.. UPSProvider
ICarrierProvider <|.. FedExProvider
ICarrierProvider <|.. USPSProvider
ICarrierProvider <|.. AmazonShippingProvider
Interface Methods:
| Method | Input | Output | Description |
|---|---|---|---|
GetRates | Origin address, destination address, package dimensions/weight | ShippingRate[] (service level, cost, estimated days) | Returns available shipping rates for the given package |
CreateLabel | Shipment details (addresses, package, service level) | Label (label URL, tracking number, cost) | Generates a shipping label and returns tracking number |
GetTracking | Tracking number | TrackingStatus (status, location, events[]) | Returns current shipment tracking status and history |
CancelShipment | Shipment ID | bool (success/failure) | Cancels a shipment and voids the label |
ValidateAddress | Address (street, city, state, zip, country) | AddressValidation (is_valid, suggested_address, corrections[]) | Validates and standardizes a shipping address |
6.10.2 Supported Carriers (Future)
All carrier integrations are planned for v2.0. The integration framework is designed to support the following carriers.
| Carrier | API | Rate Lookup | Label Generation | Tracking | Update Method | Status |
|---|---|---|---|---|---|---|
| UPS | UPS Developer API | Yes | Yes | Yes | Webhook | Planned v2.0 |
| FedEx | FedEx API | Yes | Yes | Yes | Webhook | Planned v2.0 |
| USPS | USPS Web Tools | Yes | Yes | Yes | Poll (hourly) | Planned v2.0 |
| Amazon Buy Shipping | SP-API Shipping | Yes | Yes | Yes | Push (SQS) | Planned v2.0 |
Carrier Configuration Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
carrier_code | Enum | Yes | UPS, FEDEX, USPS, AMAZON_SHIPPING |
display_name | String(100) | Yes | Admin-friendly name (e.g., “UPS Ground”) |
api_key_encrypted | String(500) | Yes | AES-256 encrypted API credentials |
api_secret_encrypted | String(500) | No | AES-256 encrypted API secret |
account_number | String(50) | Yes | Carrier account number |
is_active | Boolean | Yes | Whether this carrier is available for label generation (default: false) |
default_service_level | String(50) | No | Default shipping service (e.g., GROUND, EXPRESS, PRIORITY) |
config_json | JSON | No | Carrier-specific settings (e.g., insurance defaults, signature requirements) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
6.10.3 Shipping Data Model
Shipment Table
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table |
order_id | UUID | Yes | FK to orders table – the sales order being shipped |
carrier_code | Enum | Yes | UPS, FEDEX, USPS, AMAZON_SHIPPING |
service_level | String(50) | Yes | Service level selected (e.g., GROUND, 2DAY, OVERNIGHT) |
tracking_number | String(100) | No | Carrier tracking number (populated after label creation) |
label_url | String(500) | No | URL to printable shipping label (populated after label creation) |
ship_date | Date | No | Actual ship date (populated when carrier scans package) |
estimated_delivery | Date | No | Carrier-estimated delivery date |
actual_delivery | DateTime | No | Confirmed delivery date and time |
shipping_cost | Decimal(10,2) | Yes | Cost charged to customer for shipping |
carrier_cost | Decimal(10,2) | No | Actual cost charged by carrier (for margin reporting) |
insurance_amount | Decimal(10,2) | No | Declared value for insurance (default: 0.00) |
weight_oz | Decimal(8,2) | Yes | Package weight in ounces |
dimensions_json | JSON | No | Package dimensions: { "length": 12, "width": 8, "height": 4, "unit": "in" } |
ship_from_location_id | UUID | Yes | FK to locations table – store fulfilling the shipment |
ship_to_address_json | JSON | Yes | Destination address: { "name", "street1", "street2", "city", "state", "zip", "country" } |
status | Enum | Yes | Shipment lifecycle status (see table below) |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Shipment Status Values
| Status | Description | Trigger |
|---|---|---|
PENDING | Shipment record created, no label yet | Order marked for shipping |
LABEL_CREATED | Shipping label generated | CreateLabel API call succeeds |
SHIPPED | Package handed to carrier | Carrier first scan event |
IN_TRANSIT | Package moving through carrier network | Carrier tracking update |
OUT_FOR_DELIVERY | Package on delivery vehicle | Carrier tracking update |
DELIVERED | Package delivered to recipient | Carrier delivery confirmation |
EXCEPTION | Delivery issue (weather, address error, refused) | Carrier exception event |
RETURNED | Package returned to sender | Carrier return scan |
CANCELLED | Shipment cancelled before carrier pickup | CancelShipment API call |
6.11 Integration Hub (Enhanced)
Scope: Central configuration and health monitoring for all external system integrations. This section enhances the Integration Hub defined in Module 5, Section 5.16 by expanding the integration registry to include Amazon SP-API, Google Merchant Center, and shipping carriers alongside the existing Shopify, payment processor, and email provider integrations.
Cross-Reference: See Module 5, Section 5.16 for the base Integration Hub definition, credential storage, and Shopify-specific configuration. See Module 3, Section 3.7 for Shopify product sync logic. See Module 4, Section 4.14 for Shopify inventory sync.
6.11.1 Integration Registry (Enhanced)
The integration_type enum is expanded to include all platform integrations defined in Module 6.
Enhanced integration_type Enum
| integration_type Enum Value | Description | Status |
|---|---|---|
SHOPIFY | Shopify e-commerce platform (product, inventory, order sync) | v1.0 |
AMAZON | Amazon Selling Partner API (listings, orders, FBA/FBM) | Planned v1.5 |
GOOGLE_MERCHANT | Google Merchant Center (product feeds, local inventory) | Planned v1.5 |
PAYMENT_PROCESSOR | Card and financing processor (authorization, settlement) | v1.0 |
EMAIL_PROVIDER | Email delivery service (SMTP, SendGrid, Mailgun) | v1.0 |
SHIPPING_CARRIER | Carrier rate lookup, label generation, tracking | Planned v2.0 |
ACCOUNTING | QuickBooks Online, Xero (journal entries, invoice sync) | Planned v2.0 |
CUSTOM | Custom webhook integration (tenant-defined endpoints) | Planned v2.0 |
Integration Data Model (Enhanced)
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
tenant_id | UUID | Yes | FK to tenants table – owning tenant |
integration_type | Enum | Yes | Enhanced enum from table above |
provider_name | String(100) | Yes | Provider identifier (e.g., “Shopify”, “Amazon SP-API”, “Google Merchant Center”) |
display_name | String(100) | Yes | Admin-friendly name (e.g., “Nexus Clothing Shopify Store”, “Nexus Amazon US”) |
status | Enum | Yes | CONNECTED, DISCONNECTED, ERROR, NOT_CONFIGURED, RATE_LIMITED |
is_enabled | Boolean | Yes | Whether the integration is actively processing (default: false until verified) |
last_sync_at | DateTime | No | Timestamp of the most recent successful sync operation |
last_error_at | DateTime | No | Timestamp of the most recent error |
last_error_message | String(500) | No | Human-readable error description |
error_count_24h | Integer | Yes | Rolling count of errors in the past 24 hours (default: 0) |
sync_latency_ms | Integer | No | Average sync latency in milliseconds over the last hour |
api_version | String(20) | No | Current API version in use (e.g., “2024-01”, “2022-04-01”) |
rate_limit_remaining | Integer | No | Remaining API calls before rate limit is hit |
rate_limit_reset_at | DateTime | No | Timestamp when rate limit bucket resets |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
New fields vs. Module 5 base model: RATE_LIMITED status, api_version, rate_limit_remaining, rate_limit_reset_at.
6.11.2 Credentials Storage
Credentials for each integration are stored separately with encryption at rest. This model is identical to Module 5, Section 5.16.2.
Integration Credentials Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
integration_id | UUID | Yes | FK to integrations table |
credential_key | String(100) | Yes | Credential identifier (e.g., api_key, api_secret, access_token, refresh_token, shop_url, merchant_id, seller_id) |
credential_value_encrypted | String(1000) | Yes | AES-256 encrypted credential value. Never returned in API responses. |
expires_at | DateTime | No | Credential expiry (e.g., OAuth token, SP-API refresh token). NULL for non-expiring credentials. |
created_at | DateTime | Yes | Record creation timestamp |
updated_at | DateTime | Yes | Last modification timestamp |
Credential Rotation:
- OAuth tokens with
expires_atare automatically refreshed before expiry (5-minute buffer). - Failed token refresh triggers a
DISCONNECTEDstatus and dashboard alert. - Manual credential update requires re-verification handshake before status returns to
CONNECTED.
6.11.3 Sync Log (Enhanced)
The sync log captures every sync operation across all integrations. The sync_type enum is expanded to include new operation types introduced by Amazon (SQS notifications, bulk feeds) and Google Merchant (scheduled product pushes).
Enhanced sync_type Enum
| sync_type Value | Description | Example |
|---|---|---|
WEBHOOK_IN | Incoming webhook from external system | Shopify orders/create webhook |
WEBHOOK_OUT | Outgoing webhook to external system | POS notifies custom endpoint of sale |
SCHEDULED_PULL | Scheduled data pull from external system | Amazon order poll every 120 seconds |
SCHEDULED_PUSH | Scheduled data push to external system | Google Merchant product feed update (2x daily) |
RECONCILIATION | Periodic full data comparison and correction | Shopify inventory reconciliation every 15 minutes |
MANUAL | Admin-triggered manual sync from Integration Hub | Admin clicks “Sync Now” for Shopify products |
BULK_OPERATION | Large-scale data operation via bulk API | Shopify bulk GraphQL operation, Amazon flat-file feed |
NOTIFICATION | Asynchronous notification processing | Amazon SQS message, Google Merchant disapproval alert |
Integration Sync Log Data Model
| Field | Type | Required | Description |
|---|---|---|---|
id | UUID | Yes | Primary key |
integration_id | UUID | Yes | FK to integrations table |
sync_type | Enum | Yes | Enhanced enum from table above |
direction | Enum | Yes | INBOUND, OUTBOUND |
entity_type | String(50) | Yes | Entity synced (e.g., PRODUCT, INVENTORY, ORDER, CUSTOMER, LISTING, FEED) |
entity_id | UUID | No | FK to the local entity record affected (NULL for bulk syncs) |
external_id | String(100) | No | External system identifier (e.g., Shopify product ID, Amazon ASIN, Google offer ID) |
status | Enum | Yes | SUCCESS, FAILED, PARTIAL, SKIPPED, RETRYING |
records_processed | Integer | Yes | Number of records processed in this sync operation |
records_failed | Integer | Yes | Number of records that failed processing |
error_details | Text | No | Detailed error message and stack trace for failed syncs |
duration_ms | Integer | Yes | Sync operation duration in milliseconds |
started_at | DateTime | Yes | Sync start timestamp |
completed_at | DateTime | Yes | Sync completion timestamp |
New fields vs. Module 5 base model: RETRYING status, LISTING and FEED entity types, BULK_OPERATION and NOTIFICATION sync types.
6.11.4 Health Dashboard (Enhanced)
The Integration Hub dashboard provides a consolidated health view for all configured integrations across all six integration types.
flowchart LR
HUB["Integration Hub\n(Central Management)"]
SHOP["Shopify\nProduct + Inventory\n+ Orders"]
AMZN["Amazon SP-API\nListings + Orders\n+ FBA/FBM"]
GOOG["Google Merchant\nProduct Feeds\n+ Local Inventory"]
PAY["Payment Processor\nCard Processing\n+ Batch Settlement"]
EMAIL["Email Provider\nSMTP / SendGrid\n/ Mailgun"]
SHIP["Shipping Carrier\nRates + Labels\n+ Tracking"]
HUB -->|"Product sync\nInventory sync\nOrder webhooks"| SHOP
HUB -->|"Listings sync\nOrder poll\nSQS notifications"| AMZN
HUB -->|"Product feed\nLocal inventory\nDisapproval alerts"| GOOG
HUB -->|"Authorization\nSettlement\nRefunds"| PAY
HUB -->|"Transactional emails\nDigest notifications"| EMAIL
HUB -.->|"Rate lookup\nLabel generation\nTracking updates"| SHIP
style HUB fill:#7b2d8e,stroke:#5a1d6e,color:#fff
style SHOP fill:#2d6a4f,stroke:#1b4332,color:#fff
style AMZN fill:#ff9900,stroke:#cc7a00,color:#000
style GOOG fill:#4285f4,stroke:#3367d6,color:#fff
style PAY fill:#2d6a4f,stroke:#1b4332,color:#fff
style EMAIL fill:#2d6a4f,stroke:#1b4332,color:#fff
style SHIP fill:#6c757d,stroke:#495057,color:#fff
Enhanced Health Indicators per Integration
| Metric | Green | Yellow | Red |
|---|---|---|---|
| Status | CONNECTED | RATE_LIMITED | DISCONNECTED / ERROR |
| Error Count (24h) | 0 | 1-5 | > 5 |
| Last Sync | < 30 min ago | 30 min – 2 hrs | > 2 hrs |
| Sync Latency | < 2,000ms | 2,000 – 5,000ms | > 5,000ms |
| Rate Limit | > 50% remaining | 10-50% remaining | < 10% remaining |
| Validation Failures | 0 products | 1-10 products | > 10 products |
Dashboard Features:
- Real-time status indicator (green/yellow/red dot) per integration.
- Click-through to detailed sync log filtered by integration.
- “Sync Now” button for manual sync trigger (requires ADMIN role).
- Credential expiry warning banner (30 days before OAuth token expiry).
- Rate limit usage bar showing current consumption vs. limit.
6.12 Integration Business Rules (YAML)
Scope: Consolidated YAML configuration for all integration-related business rules across Module 6. This follows the same pattern as Module 5, Section 5.19 (Consolidated Business Rules). All values shown are defaults and can be overridden at tenant level.
# ============================================================
# Module 6: Integration Business Rules
# ============================================================
# All values shown are defaults and can be overridden at
# tenant level. This file is the single authoritative source
# for all integration-related configurable settings.
# ============================================================
integration_config:
# ----------------------------------------------------------
# SHOPIFY INTEGRATION
# ----------------------------------------------------------
shopify:
sync_mode: "pos_master" # pos_master | bidirectional
api_preference: "graphql" # graphql | rest
idempotency_required: true
third_party_pos: true # Shopify POS channel disabled; POS is external
source_of_truth: "pos" # pos | shopify (for product data)
track_inventory: true # Shopify inventory tracking enabled
bopis_enabled: true # Buy Online, Pick Up In Store
reconciliation_interval_minutes: 15
webhook_hmac_algorithm: "sha256"
max_variants_per_product: 100 # Shopify hard limit
max_option_dimensions: 3 # Shopify hard limit
bulk_concurrent_queries: 5
bulk_concurrent_mutations: 5
rate_limit_rest_bucket: 40 # Shopify REST leak bucket size
rate_limit_rest_leak_per_sec: 2 # Shopify REST leak rate
rate_limit_graphql_points_per_sec: 50
image_sync: "first_publish_only" # first_publish_only | always | never
customer_sync_enabled: true
order_sync_enabled: true
# ----------------------------------------------------------
# AMAZON SP-API INTEGRATION
# ----------------------------------------------------------
amazon:
enabled: false # Disabled by default; tenant must opt-in
sync_mode: "pos_master" # pos_master only (Amazon does not push product edits)
marketplace_id: "ATVPDKIKX0DER" # US marketplace (configurable per tenant)
region: "NA" # NA | EU | FE
order_poll_interval_seconds: 120 # Poll for new orders every 2 minutes
fba_enabled: false # Fulfillment by Amazon
fbm_enabled: true # Fulfillment by Merchant (store ships)
safety_buffer_default_qty: 10 # Reserve 10 units from Amazon availability
safety_buffer_default_pct: null # Percentage-based buffer (overrides qty if set)
notification_delivery: "sqs" # sqs | polling
catalog_api_version: "2022-04-01"
listings_api_version: "2021-08-01"
packaging_label_format: "4x6" # Thermal label format
seller_code_compliance: true # Enforce Amazon seller code rules
max_bullet_points: 5 # Amazon listing bullet point limit
max_bullet_point_length: 1000 # Characters per bullet point
max_search_terms_bytes: 250 # Amazon search term byte limit
fulfillment_default: "FBM" # FBM | FBA
# ----------------------------------------------------------
# GOOGLE MERCHANT CENTER INTEGRATION
# ----------------------------------------------------------
google_merchant:
enabled: false # Disabled by default; tenant must opt-in
sync_mode: "pos_master" # POS is source of truth for product data
api_version: "v1" # Content API version
local_inventory_enabled: false # Local Inventory Ads (LIA)
product_update_frequency: "2x_daily" # 2x_daily | daily | hourly
image_validation_strict: true # Enforce Google image requirements
gtin_required: true # GTIN (barcode) required for all products
price_match_validation: true # Verify POS price matches Google listing
disapproval_prevention: true # Pre-validate before pushing to Google
ssl_required: true # Landing page must be HTTPS
content_api_migration_deadline: "2026-08-18" # Merchant API migration deadline
v1beta_deadline: "2026-02-28" # v1beta sunset date
title_max_length: 150 # Google product title limit
description_max_length: 5000 # Google product description limit
gbp_integration_enabled: false # Google Business Profile integration
# ----------------------------------------------------------
# CROSS-PLATFORM PRODUCT VALIDATION
# ----------------------------------------------------------
product_validation:
title_max_length: 150 # Strictest platform limit
description_max_length: 5000 # Strictest platform limit
image_min_width: 1000 # Pixels (Google requirement)
image_min_height: 1000 # Pixels (Google requirement)
image_max_size_mb: 10
image_formats_allowed:
- "JPEG"
- "PNG"
watermarks_prohibited: true # Google and Amazon both prohibit
text_overlay_prohibited: true # Google prohibits text on images
white_background_required: true # Amazon main image requirement
barcode_required: true # GTIN/UPC/EAN required for marketplace listing
brand_required: true # Required by Google and Amazon
condition_required: true # Required by Google
condition_default: "new" # new | refurbished | used
sku_max_length: 50
weight_required: true # Required for shipping calculation
product_type_required: true # Google product category taxonomy
# ----------------------------------------------------------
# CROSS-PLATFORM INVENTORY SYNC
# ----------------------------------------------------------
inventory_sync:
safety_buffer_enabled: true # Hold back inventory from marketplaces
safety_buffer_default_qty: 0 # Default buffer (0 = no buffer)
oversell_prevention: true # Block sale if available_qty <= 0
reserve_on_order: true # Reserve inventory when order is placed
first_commit_wins: true # First system to commit gets the unit
sync_failure_freeze_minutes: 120 # Freeze marketplace qty on sync failure
dead_letter_retry_hours: 24 # Retry failed sync events for 24 hours
shopify_reconciliation_minutes: 15
amazon_reconciliation_minutes: 30
google_reconciliation_hours: 6
# ----------------------------------------------------------
# PAYMENT INTEGRATION
# ----------------------------------------------------------
payment:
pci_scope: "SAQ-A" # Semi-integrated; card data never touches POS
payment_timeout_seconds: 60
connection_timeout_seconds: 10
same_day_void: true
batch_close_time: "23:00" # Local time, configurable per location
batch_close_auto: true
terminal_failure_alert_threshold: 3 # Consecutive failures before alert
terminal_failure_alert_window_min: 15
reconciliation_variance_tolerance: 0.01 # Dollars
decline_log_retention_days: 90
# ----------------------------------------------------------
# EMAIL INTEGRATION
# ----------------------------------------------------------
email:
provider_type: "SMTP" # SMTP | SENDGRID | MAILGUN
daily_send_limit: 0 # 0 = unlimited
bounce_rate_alert_threshold_pct: 5
consecutive_failure_suppress: 3 # Suppress after N failures to same address
delivery_log_retention_days: 90
test_email_required_before_go_live: true
# ----------------------------------------------------------
# SHIPPING INTEGRATION (Future)
# ----------------------------------------------------------
shipping:
enabled: false # Planned for v2.0
default_carrier: null
address_validation_required: true
insurance_default: false
label_format: "4x6" # Thermal label format
tracking_poll_interval_minutes: 60
# ----------------------------------------------------------
# GLOBAL INTEGRATION SETTINGS
# ----------------------------------------------------------
global:
retry_max_attempts: 3
retry_backoff_base_seconds: 5
retry_backoff_multiplier: 3 # Exponential: 5s, 15s, 45s
circuit_breaker_threshold: 5 # Failures before circuit opens
circuit_breaker_window_seconds: 60
circuit_breaker_cooldown_seconds: 30
idempotency_window_hours: 24 # Idempotency key validity
credential_encryption: "AES-256"
webhook_verification: "HMAC-SHA256"
sync_log_retention_days: 90
health_check_interval_seconds: 60
rate_limit_buffer_pct: 10 # Stop at 90% of rate limit
6.13 Integration User Stories & Gherkin Acceptance Criteria
Scope: All user stories and Gherkin acceptance criteria for Module 6 (Integrations). Stories are organized into 7 epics covering all integration areas. Each epic includes user stories in standard format and Gherkin feature files with acceptance scenarios.
6.13.1 Integration User Story Epics
Epic 6.A: Shopify Integration
US-6.A.1: Product Sync to Shopify
- As a store manager, I want products created in POS to automatically appear on my Shopify store so that my online catalog stays current without manual duplicate entry.
- Constraint: Product must have all required fields (title, price, at least one image, barcode). Sync occurs within 30 seconds of product creation.
US-6.A.2: Real-Time Inventory Sync
- As a store manager, I want real-time inventory sync between POS and Shopify so that online customers never purchase items that are out of stock in-store.
- Constraint: Inventory updates propagate within 30 seconds. Reconciliation runs every 15 minutes.
US-6.A.3: Online Order Fulfillment
- As a store manager, I want online Shopify orders to appear in my POS for fulfillment so that my staff can pick, pack, and ship from the store.
- Constraint: Orders appear within 60 seconds of placement. Inventory is reserved immediately.
US-6.A.4: Sync Mode Configuration
- As a tenant admin, I want to configure Shopify sync mode (POS-master or bidirectional) so that I can control which system is authoritative for product data.
- Constraint: Inventory sync is always bidirectional regardless of product sync mode.
US-6.A.5: BOPIS Order Processing
- As a store manager, I want BOPIS (Buy Online, Pick Up In Store) orders from Shopify to appear as pickup orders in POS so that staff can stage items for customer collection.
- Constraint: BOPIS orders follow the Hold for Pickup workflow (Module 1, Section 1.11).
US-6.A.6: Shopify Conflict Resolution
- As a tenant admin, I want the system to automatically resolve sync conflicts between POS and Shopify so that data remains consistent without manual intervention.
- Constraint: In POS-master mode, POS-owned field changes in Shopify are overwritten on next sync cycle.
Epic 6.B: Amazon SP-API Integration
US-6.B.1: Amazon Account Connection
- As a tenant admin, I want to connect my Amazon Seller Central account to POS via OAuth so that I can manage Amazon listings from within the POS system.
- Constraint: Connection uses SP-API OAuth with LWA (Login with Amazon). Refresh token stored encrypted.
US-6.B.2: Amazon Listing Management
- As a store manager, I want products listed on Amazon to be managed from POS so that pricing, descriptions, and images are maintained in one place.
- Constraint: POS is source of truth. Amazon-specific fields (bullet points, search terms) are editable from POS.
US-6.B.3: Amazon FBM Order Routing
- As a store manager, I want Amazon FBM orders routed to the nearest store for fulfillment so that delivery times are minimized and inventory is balanced.
- Constraint: Store assignment uses the same proximity + stock algorithm as Shopify orders (Module 4, Section 4.14).
US-6.B.4: Safety Buffer Configuration
- As a tenant admin, I want to configure safety buffers for Amazon inventory so that I can reserve stock for in-store customers and prevent overselling.
- Constraint: Buffer can be absolute quantity or percentage. Applied per-product or globally.
US-6.B.5: FBA Inventory Monitoring
- As a store manager, I want to monitor FBA inventory levels from within POS so that I can see total inventory across all fulfillment channels.
- Constraint: FBA quantities are read-only in POS. Displayed separately from FBM/in-store quantities.
US-6.B.6: Amazon Order Polling
- As a tenant admin, I want the system to automatically poll for new Amazon orders so that orders are imported without manual action.
- Constraint: Polling interval configurable (default: 120 seconds). SQS notifications preferred when available.
Epic 6.C: Google Merchant Integration
US-6.C.1: Google Merchant Connection
- As a tenant admin, I want to connect Google Merchant Center to POS via OAuth so that product data feeds to Google Shopping automatically.
- Constraint: Uses Google Content API for Shopping. Service account or OAuth credentials stored encrypted.
US-6.C.2: Local Inventory Ads
- As a store manager, I want local store inventory to appear in Google Shopping searches so that nearby customers can see what is available at my store.
- Constraint: Requires Google Business Profile linked to Merchant Center. Location-level inventory synced.
US-6.C.3: Pre-Publish Validation
- As a tenant admin, I want products validated against Google requirements before sync to prevent disapprovals so that my product listings maintain good standing.
- Constraint: Validation checks GTIN, images (1000x1000 min, no watermarks), price match, and required attributes.
US-6.C.4: Disapproval Dashboard
- As a store manager, I want a disapproval dashboard showing which products failed Google validation so that I can fix issues before they affect visibility.
- Constraint: Dashboard shows disapproval reason, affected product, date, and suggested fix.
US-6.C.5: Product Feed Scheduling
- As a tenant admin, I want to configure product feed update frequency so that Google always has current product data without excessive API usage.
- Constraint: Options are hourly, daily, or 2x daily (default). Changes trigger immediate incremental sync.
US-6.C.6: Price Consistency Enforcement
- As a tenant admin, I want the system to verify that POS prices match Google listing prices so that customers are not misled by stale pricing.
- Constraint: Price mismatch detected during reconciliation triggers an automatic price update push.
Epic 6.D: Cross-Platform Product Validation
US-6.D.1: Unified Validation Engine
- As a product manager, I want products validated against all platform requirements before publishing so that I can fix issues once rather than per-platform.
- Constraint: Validation runs against the strictest requirement across all enabled platforms.
US-6.D.2: Image Validation
- As a product manager, I want image validation that checks dimensions, file size, watermarks, and background requirements so that images pass all platform reviews.
- Constraint: Minimum 1000x1000px, max 10MB, JPEG/PNG only, no watermarks or text overlays, white background for Amazon main image.
US-6.D.3: Unified Validation Dashboard
- As a product manager, I want a unified validation dashboard showing platform readiness per product so that I can see at a glance which products are ready for each channel.
- Constraint: Dashboard shows green/yellow/red status per product per platform with drill-down to specific failures.
US-6.D.4: Validation on Product Save
- As a product manager, I want validation to run automatically when I save a product so that I receive immediate feedback on any issues.
- Constraint: Validation is non-blocking (product saves regardless) but warnings are displayed prominently.
US-6.D.5: Bulk Validation Report
- As a tenant admin, I want to run bulk validation across all products so that I can identify and fix issues before enabling a new marketplace.
- Constraint: Bulk validation runs asynchronously and produces a downloadable report.
Epic 6.E: Cross-Platform Inventory Sync
US-6.E.1: Safety Buffer Management
- As a tenant admin, I want to configure safety buffers per marketplace so that I can reserve stock for in-store customers while selling online.
- Constraint: Buffers are configurable per product, per marketplace. Default is configurable at tenant level.
US-6.E.2: Oversell Prevention
- As a store manager, I want the system to prevent overselling across all channels so that customers never purchase items that are not available.
- Constraint: First-commit-wins arbitration. Inventory reserved at time of order, not payment.
US-6.E.3: Sync Failure Handling
- As a tenant admin, I want inventory quantities frozen on marketplaces when sync fails so that stale data does not cause overselling.
- Constraint: Quantities are frozen for configurable period (default: 120 minutes). Dead letter queue retries for 24 hours.
US-6.E.4: Reconciliation Dashboard
- As a store manager, I want a reconciliation dashboard showing inventory discrepancies across channels so that I can identify and resolve sync issues.
- Constraint: Dashboard shows POS qty vs. each marketplace qty with variance highlighting.
US-6.E.5: Multi-Channel Available Quantity
- As a store manager, I want to see available quantity per channel from the product detail screen so that I understand how inventory is allocated.
- Constraint: Displays: Total On Hand, Shopify Available, Amazon Available, Google Available, In-Store Reserve, Safety Buffer.
Epic 6.F: Payment Integration
US-6.F.1: Card Payment Processing
- As a cashier, I want to process card payments via the payment terminal without handling card data so that transactions are fast and PCI compliant.
- Constraint: Card data never enters POS system. Terminal communicates directly with processor. POS stores only token and masked card.
US-6.F.2: Processor Credential Configuration
- As a tenant admin, I want to configure payment processor credentials with test/production toggle so that I can verify the integration in sandbox before going live.
- Constraint: Credentials encrypted with AES-256. Validation handshake required before activating production mode.
US-6.F.3: Terminal Health Monitoring
- As a store manager, I want to view payment terminal health and decline rates so that I can identify and resolve terminal issues proactively.
- Constraint: Dashboard shows per-terminal metrics: transaction count, avg response time, error rate, decline rate.
US-6.F.4: Batch Settlement Management
- As a store manager, I want to view daily batch settlement reports and reconciliation status so that I can verify all card transactions settled correctly.
- Constraint: Auto-batch at configurable time. Variance > $0.01 triggers reconciliation alert.
Epic 6.G: Integration Hub
US-6.G.1: Integration Management
- As a tenant admin, I want a central dashboard to manage all external integrations so that I can connect, configure, and monitor all third-party services from one place.
- Constraint: Dashboard shows status, last sync, error count, and health indicator per integration.
US-6.G.2: Health Monitoring
- As a tenant admin, I want real-time health indicators for all integrations so that I can immediately see when an integration needs attention.
- Constraint: Green/yellow/red indicators based on status, error count, sync age, latency, and rate limit usage.
US-6.G.3: Sync Log Access
- As a tenant admin, I want to view detailed sync logs for each integration so that I can troubleshoot failed syncs and understand data flow.
- Constraint: Logs filterable by integration, sync type, status, date range. 90-day retention.
US-6.G.4: Manual Sync Trigger
- As a tenant admin, I want to manually trigger a sync for any integration so that I can force a refresh when needed without waiting for the scheduled cycle.
- Constraint: Requires ADMIN or OWNER role. Rate limited to one manual sync per integration per 5 minutes.
6.13.2 Gherkin Acceptance Criteria
Feature: Shopify Product Sync
As a store manager
I want products to sync between POS and Shopify
So that my online store always shows current product data
Background:
Given I am logged in as a user with "MANAGER" role
And Shopify integration is enabled with status "CONNECTED"
And sync_mode is set to "pos_master"
Scenario: New product syncs to Shopify on creation
Given I create a new product with SKU "NXJ-TSHIRT-BLK-M"
And the product has title "Classic Black T-Shirt - Medium"
And the product has base_price "$24.99"
And the product has a valid image (1200x1200px JPEG)
And the product has barcode "195962000123"
When the product is saved
Then the product should appear in Shopify within 30 seconds
And the Shopify product title should match "Classic Black T-Shirt - Medium"
And the Shopify product price should match "$24.99"
And a sync log entry should be created with sync_type "WEBHOOK_OUT" and status "SUCCESS"
Scenario: POS-owned field change in Shopify is overwritten
Given product "NXJ-TSHIRT-BLK-M" exists in both POS and Shopify
And sync_mode is "pos_master"
When someone changes the Shopify title to "Updated Title on Shopify"
And the next reconciliation cycle runs
Then the Shopify title should revert to the POS title "Classic Black T-Shirt - Medium"
And a sync log entry should be created with sync_type "RECONCILIATION"
Scenario: Product missing required fields is excluded from sync
Given I create a new product with SKU "NXJ-DRAFT-001"
And the product has title "Draft Product"
But the product has no image
When the product is saved
Then the product should NOT sync to Shopify
And a validation warning should display "Product excluded from Shopify sync: missing required image"
And the product should be saved locally without error
Feature: Shopify Inventory Sync
As a store manager
I want real-time inventory sync between POS and Shopify
So that online customers see accurate stock levels
Background:
Given Shopify integration is enabled with inventory_sync_enabled = true
And product "NXJ-TSHIRT-BLK-M" exists in both POS and Shopify
And the current POS quantity at "Georgetown Store" is 25
Scenario: POS sale decrements Shopify inventory
When a sale of 1 unit of "NXJ-TSHIRT-BLK-M" is completed at "Georgetown Store"
Then the POS quantity should update to 24
And the Shopify inventory level should update to 24 within 30 seconds
And a sync log entry should be created with entity_type "INVENTORY" and status "SUCCESS"
Scenario: Shopify order decrements POS inventory
When a Shopify order for 2 units of "NXJ-TSHIRT-BLK-M" is placed
Then the POS should receive the order via webhook within 60 seconds
And the POS quantity should decrease by 2 (from 25 to 23)
And the inventory status should show 2 units as "RESERVED" for the online order
Scenario: Reconciliation detects and corrects discrepancy
Given the POS quantity is 20 but the Shopify quantity shows 22
When the scheduled reconciliation runs (every 15 minutes)
Then the Shopify quantity should be corrected to 20 (POS is source of truth)
And a sync log entry should be created with sync_type "RECONCILIATION"
And the discrepancy should be logged with details "Corrected Shopify qty from 22 to 20"
Scenario: Inventory sync failure freezes marketplace quantity
Given the Shopify API is unreachable
When a POS sale occurs reducing quantity from 20 to 19
Then the POS should record the sync as "FAILED"
And the Shopify quantity should remain frozen at its last known value
And a retry should be queued with exponential backoff (5s, 15s, 45s)
And if still failing after 120 minutes, an alert should be sent to the tenant admin
Feature: Amazon Order Import
As a store manager
I want Amazon orders automatically imported into POS
So that I can fulfill them from my store inventory
Background:
Given Amazon SP-API integration is enabled with status "CONNECTED"
And fulfillment_default is "FBM"
And order_poll_interval_seconds is 120
Scenario: New FBM order is imported via polling
Given a customer places an Amazon order for "NXJ-TSHIRT-BLK-M" (qty: 1)
When the next order poll cycle runs
Then the order should appear in POS with source "AMAZON"
And the order status should be "PENDING_FULFILLMENT"
And inventory should be reserved (1 unit) at the assigned store
And a sync log entry should be created with sync_type "SCHEDULED_PULL" and entity_type "ORDER"
Scenario: FBM order is routed to nearest store with stock
Given "Georgetown Store" has 5 units and "HQ Warehouse" has 20 units
And the customer shipping address is in Washington, DC
When an Amazon FBM order is imported
Then the order should be assigned to "Georgetown Store" (nearest with stock)
And the fulfillment team at "Georgetown Store" should see the order in their queue
Scenario: Order for out-of-stock item triggers alert
Given "NXJ-TSHIRT-BLK-M" has 0 units available across all locations
When an Amazon order for this item is imported
Then the order should be created with status "PENDING_FULFILLMENT"
And a "STOCK_ALERT" notification should be sent to the store manager
And the order should be flagged with warning "Insufficient stock for fulfillment"
Feature: Amazon FBM Fulfillment
As a store manager
I want to fulfill Amazon FBM orders from my store
So that customers receive their orders on time
Background:
Given Amazon integration is connected
And I am logged in as a user with "MANAGER" role at "Georgetown Store"
Scenario: Ship and confirm FBM order
Given an Amazon FBM order "114-1234567-8901234" is assigned to my store
And the order contains 1 unit of "NXJ-TSHIRT-BLK-M"
When I pick and pack the item
And I generate a shipping label via the carrier integration
And I confirm shipment with tracking number "1Z999AA10123456784"
Then the order status should update to "SHIPPED" in POS
And Amazon should receive the shipment confirmation via SP-API
And the customer should receive a shipping notification from Amazon
And the inventory should be decremented by 1 unit at "Georgetown Store"
Scenario: FBA inventory is visible but read-only
Given FBA is enabled for product "NXJ-HOODIE-GRY-L"
And Amazon FBA has 50 units in their fulfillment center
When I view the product detail for "NXJ-HOODIE-GRY-L"
Then I should see "FBA Qty: 50" in the inventory breakdown
And the FBA quantity should be read-only (not editable from POS)
And the total network quantity should include FBA units
Feature: Google Local Inventory
As a store manager
I want local store inventory visible in Google Shopping
So that nearby customers can find products at my store
Background:
Given Google Merchant integration is connected
And local_inventory_enabled is true
And Google Business Profile is linked for "Georgetown Store"
Scenario: Local inventory appears in Google Shopping
Given "NXJ-TSHIRT-BLK-M" has 15 units at "Georgetown Store"
And the product passes all Google validation rules
When the scheduled product feed runs
Then Google Merchant should show "In stock" at "Georgetown Store"
And the price shown should match the POS base_price "$24.99"
Scenario: Out-of-stock local inventory updates Google
Given "NXJ-TSHIRT-BLK-M" has 0 units at "Georgetown Store"
When the next inventory sync to Google runs
Then Google Merchant should show "Out of stock" at "Georgetown Store"
And other locations with stock should still show "In stock"
Scenario: Product with missing GTIN is excluded from Google feed
Given product "NXJ-CUSTOM-001" has no barcode/GTIN
And gtin_required is true
When the product feed sync runs
Then "NXJ-CUSTOM-001" should be excluded from the Google feed
And a validation failure should be logged with reason "Missing required GTIN"
And the disapproval dashboard should show this product
Feature: Cross-Platform Product Validation
As a product manager
I want products validated against all platform requirements
So that I can publish to any channel with confidence
Background:
Given Shopify, Amazon, and Google Merchant integrations are all enabled
And I am logged in as a user with "PRODUCT_MANAGER" role
Scenario: Product passes all platform validations
Given product "NXJ-TSHIRT-BLK-M" has:
| Field | Value |
| title | Classic Black T-Shirt - Medium |
| description | Premium cotton crew neck t-shirt |
| price | $24.99 |
| barcode | 195962000123 |
| brand | Nexus Clothing |
| weight | 8 oz |
| image | 1200x1200 JPEG, no watermark, white background |
| condition | new |
When I save the product
Then the validation dashboard should show:
| Platform | Status |
| Shopify | Ready (green) |
| Amazon | Ready (green) |
| Google | Ready (green) |
Scenario: Product fails Amazon image validation
Given product "NXJ-DRESS-RED-S" has a main image with a colored background
And white_background_required is true for Amazon
When I save the product
Then the validation dashboard should show:
| Platform | Status |
| Shopify | Ready (green) |
| Amazon | Warning (yellow) - "Main image requires white background" |
| Google | Warning (yellow) - "Image may not meet quality standards" |
Scenario: Product exceeds Shopify variant limit
Given product "NXJ-MATRIX-SHOE" has 120 variants (sizes x colors x widths)
And max_variants_per_product is 100
When I save the product
Then the validation dashboard should show Shopify status as "Blocked (red)"
And the message should read "Exceeds Shopify 100-variant limit (120 variants)"
And the product should be excluded from Shopify sync
But the product should still be eligible for Amazon and Google sync
Scenario: Bulk validation report generation
Given there are 500 active products in the catalog
When I click "Run Bulk Validation" from the validation dashboard
Then a background job should start processing all 500 products
And I should see a progress indicator
And when complete, a downloadable CSV report should be available
And the report should contain one row per product per platform with validation status
Feature: Safety Buffer Inventory Management
As a tenant admin
I want to configure safety buffers per marketplace
So that I reserve stock for in-store customers while selling online
Background:
Given I am logged in as a user with "ADMIN" role
And product "NXJ-TSHIRT-BLK-M" has 50 total units on hand
Scenario: Safety buffer reduces marketplace available quantity
Given the Amazon safety buffer is set to 10 units
And the Shopify safety buffer is set to 0 units
When inventory sync runs for "NXJ-TSHIRT-BLK-M"
Then Shopify should show 50 available units
And Amazon should show 40 available units (50 - 10 buffer)
Scenario: Percentage-based safety buffer
Given the Amazon safety buffer is set to 20%
And total on-hand quantity is 50
When inventory sync runs
Then Amazon should show 40 available units (50 - 10 buffer, where 10 = 20% of 50)
Scenario: Safety buffer prevents overselling
Given the Amazon safety buffer is 10 units
And total on-hand is 12 units
And Amazon shows 2 available (12 - 10)
When an Amazon order for 3 units is placed
Then the order should be imported with a warning "Ordered qty (3) exceeds Amazon available (2)"
And inventory should be reserved for 3 units (allowing negative available on Amazon)
And the Amazon available quantity should update to 0
Scenario: Buffer recalculated on inventory change
Given the Amazon safety buffer is 10 units
And total on-hand is 50 (Amazon shows 40)
When a POS sale reduces on-hand to 45
Then Amazon available should update to 35 (45 - 10)
And this update should propagate within the sync interval
Feature: Payment Terminal Flow
As a cashier
I want to process card payments securely via the terminal
So that customers can pay by card without handling card data
Background:
Given I am logged in as a user with "CASHIER" role
And a payment terminal "TRM-001-GM" is configured and active at my location
And a cart with total $45.00 is ready for payment
Scenario: Successful card payment via tap
When I click "Pay by Card"
And the terminal displays "Tap or Insert Card - $45.00"
And the customer taps their Visa card
Then the terminal should communicate directly with the processor
And the POS should receive: token, approval_code "AUTH4829", masked_card "****1234", brand "VISA"
And the POS should display "Approved - $45.00"
And the receipt should show "VISA ****1234"
And no card data should be stored in the POS database
Scenario: Card declined - insufficient funds
When I click "Pay by Card"
And the customer inserts their card
And the processor returns "DECLINED - Insufficient Funds"
Then the POS should display "Card Declined: Insufficient Funds"
And I should see options: "Try Another Card" | "Cash" | "Cancel"
And the decline should be logged with reason "Insufficient Funds"
Scenario: Terminal timeout
When I click "Pay by Card"
And the terminal does not respond within 60 seconds
Then the POS should display "Terminal not responding"
And I should see options: "Retry" | "Different Terminal" | "Cash" | "Cancel"
And the timeout should be logged as a terminal failure event
Feature: Integration Health Dashboard
As a tenant admin
I want to monitor the health of all integrations
So that I can quickly identify and resolve connectivity issues
Background:
Given I am logged in as a user with "ADMIN" role
And I navigate to the Integration Hub dashboard
Scenario: All integrations healthy
Given Shopify integration last synced 5 minutes ago with 0 errors
And Payment Processor status is "CONNECTED" with 0 errors
And Email Provider is verified and has 0 bounces
When I view the Integration Hub dashboard
Then Shopify should show a green health indicator
And Payment Processor should show a green health indicator
And Email Provider should show a green health indicator
Scenario: Integration in error state shows red
Given Shopify integration has had 8 errors in the past 24 hours
And the last sync was 3 hours ago
When I view the Integration Hub dashboard
Then Shopify should show a red health indicator
And the error count should display "8 errors (24h)"
And the last sync should display "3 hours ago" in red text
And an "Investigate" button should be available
Scenario: Rate-limited integration shows yellow
Given Amazon SP-API has 8% of rate limit remaining
And the rate_limit_reset_at is in 45 seconds
When I view the Integration Hub dashboard
Then Amazon should show a yellow health indicator
And the rate limit bar should show "8% remaining"
And a tooltip should display "Rate limit resets in 45 seconds"
Feature: Integration Credential Management
As a tenant admin
I want to securely manage integration credentials
So that external services stay connected without exposing sensitive data
Background:
Given I am logged in as a user with "OWNER" role
And I navigate to the Integration Hub settings
Scenario: Configure Shopify credentials
When I click "Configure" on the Shopify integration card
And I enter shop URL "nexus-clothes.myshopify.com"
And I enter API key "shppa_abc123def456"
And I enter API secret "shpss_xyz789"
And I click "Verify & Save"
Then the system should perform a test API call to Shopify
And the credentials should be encrypted with AES-256 before storage
And the integration status should change to "CONNECTED"
And the API key should display as "shppa_****f456" in the UI
Scenario: Credentials never exposed in API response
Given Shopify integration is connected with saved credentials
When I make a GET request to /api/integrations/{shopify_id}
Then the response should contain credential_key values
But credential_value_encrypted should NOT be in the response
And a masked indicator "********" should be shown instead
Scenario: Failed credential verification
When I enter invalid Shopify API credentials
And I click "Verify & Save"
Then the system should display "Verification failed: Invalid API credentials"
And the integration status should remain "NOT_CONFIGURED"
And the invalid credentials should NOT be saved
Scenario: OAuth token auto-refresh
Given Amazon integration has a refresh token expiring in 4 minutes
When the token refresh check runs (5-minute buffer)
Then the system should automatically request a new access token
And the new token should be encrypted and stored
And the integration status should remain "CONNECTED"
And a sync log entry should record the token refresh
End of Module 6: Integrations & External Systems (Sections 6.1 – 6.13)
7. State Machine Reference
This section consolidates all entity state machines for quick reference.
7.1 Order States
| State | Description | Transitions To |
|---|---|---|
| DRAFT | Cart in progress | PENDING |
| PENDING | Awaiting payment | PAID, PARTIAL_PAID |
| PARTIAL_PAID | Partial payment received | PAID, PENDING |
| PAID | Full payment received | COMPLETED, HOLD_FOR_PICKUP |
| HOLD_FOR_PICKUP | Paid, awaiting pickup staging | READY_FOR_PICKUP |
| READY_FOR_PICKUP | Items staged for pickup | COMPLETED, HOLD_EXPIRED |
| HOLD_EXPIRED | Pickup deadline passed | CONTACT_CUSTOMER |
| CONTACT_CUSTOMER | Staff attempting contact | READY_FOR_PICKUP (extended), REFUNDED |
| COMPLETED | Transaction finalized | VOIDED (same day), PARTIALLY_RETURNED |
| VOIDED | Transaction reversed (same day) | Terminal |
| PARTIALLY_RETURNED | Some items returned | FULLY_RETURNED |
| FULLY_RETURNED | All items returned | Terminal |
| REFUNDED | Customer refunded (expired hold) | Terminal |
7.2 Parked Sale States
| State | Description | Transitions To |
|---|---|---|
| ACTIVE | Cart in progress | PARKED, PENDING |
| PARKED | Sale parked for later | ACTIVE (retrieved), EXPIRED |
| EXPIRED | TTL exceeded (4 hours) | Terminal (inventory released) |
7.3 Gift Card States
| State | Description | Transitions To |
|---|---|---|
| INACTIVE | Card manufactured, not sold | ACTIVE |
| ACTIVE | Balance > $0, within expiry | DEPLETED, EXPIRED (where allowed) |
| DEPLETED | Balance = $0 | ACTIVE (reload), CASHED_OUT |
| CASHED_OUT | Cash out processed (CA) | Terminal |
| EXPIRED | Past expiry date (where allowed) | Terminal |
7.4 Layaway States
| State | Description | Transitions To |
|---|---|---|
| DEPOSIT_PAID | Initial deposit received | RESERVED |
| RESERVED | Inventory held | PAID_IN_FULL, CANCELLED, FORFEITED |
| PAID_IN_FULL | All payments complete | COMPLETED |
| COMPLETED | Items released to customer | Terminal |
| CANCELLED | Customer cancelled | Terminal |
| FORFEITED | Payment deadline missed | Terminal |
7.5 Special Order States
| State | Description | Transitions To |
|---|---|---|
| CREATED | Order initiated | DEPOSIT_PAID, CANCELLED |
| DEPOSIT_PAID | Deposit received | ORDERED, CANCELLED_REFUND |
| ORDERED | Sent to vendor | RECEIVED |
| RECEIVED | Item arrived | READY_FOR_PICKUP |
| READY_FOR_PICKUP | Staged for customer | COMPLETED, ABANDONED |
| COMPLETED | Customer picked up | Terminal |
| CANCELLED | No deposit, cancelled | Terminal |
| CANCELLED_REFUND | Deposit refunded | Terminal |
| ABANDONED | No pickup after 30 days | Terminal |
7.6 Transfer States
| State | Description | Transitions To |
|---|---|---|
| REQUESTED | Transfer initiated | PAID, CANCELLED |
| PAID | Customer paid in full | PICKING |
| PICKING | Source store processing | SHIPPED |
| SHIPPED | Handed to carrier | IN_TRANSIT |
| IN_TRANSIT | Carrier confirmed pickup | RECEIVED |
| RECEIVED | Arrived at destination | COMPLETED |
| COMPLETED | Customer notified/picked up | Terminal |
| CANCELLED | Cancelled before payment | Terminal |
| CANCELLED_REFUND | Cancelled after payment | Terminal |
7.7 Reservation States
| State | Description | Transitions To |
|---|---|---|
| REQUESTED | Reservation initiated | PAID, CANCELLED |
| PAID | Customer paid in full | RESERVED |
| RESERVED | Item held at store | PICKED_UP, EXPIRED |
| PICKED_UP | Customer collected | Terminal |
| EXPIRED | Deadline passed | REFUND_PENDING |
| REFUND_PENDING | Auto-refund triggered | REFUNDED |
| CANCELLED | Cancelled before payment | Terminal |
| REFUNDED | Refund processed | Terminal |
7.8 Ship-to-Customer States
| State | Description | Transitions To |
|---|---|---|
| REQUESTED | Shipment initiated | PAID, CANCELLED |
| PAID | Customer paid item + shipping | PICKING, CANCELLED_REFUND |
| PICKING | Source store processing | PACKED |
| PACKED | Items packed, awaiting label | SHIPPED |
| SHIPPED | Label generated, handed to carrier | IN_TRANSIT |
| IN_TRANSIT | Carrier pickup confirmed | DELIVERED |
| DELIVERED | Delivery confirmed | Terminal |
| CANCELLED | Cancelled before payment | Terminal |
| CANCELLED_REFUND | Cancelled after payment, full refund | Terminal |
7.9 Cash Drawer States
| State | Description | Transitions To |
|---|---|---|
| CLOSED | Drawer secured | OPENING |
| OPENING | Manager initiating open | OPEN |
| OPEN | Accepting transactions | COUNTING |
| COUNTING | End-of-day count in progress | BALANCED, VARIANCE_DETECTED |
| BALANCED | Count matches expected | CLOSED |
| VARIANCE_DETECTED | Count doesn’t match | MANAGER_REVIEW |
| MANAGER_REVIEW | Awaiting approval | BALANCED |
7.10 Coupon States
| State | Description | Transitions To |
|---|---|---|
| CREATED | Coupon generated | ACTIVE |
| ACTIVE | Available for use | REDEEMED, EXPIRED, DEPLETED |
| REDEEMED | Single-use completed | Terminal |
| EXPIRED | Past expiry date | Terminal |
| DEPLETED | Multi-use limit reached | Terminal |
7.11 Customer Tier States
| State | Description | Transitions To |
|---|---|---|
| BRONZE | New/base tier | SILVER |
| SILVER | Mid tier ($1,000+ annual) | GOLD, BRONZE |
| GOLD | Top tier ($5,000+ annual) | SILVER |
7.12 Offline Mode States
| State | Description | Transitions To |
|---|---|---|
| ONLINE | Network available | OFFLINE |
| OFFLINE | Network lost, queuing locally | SYNCING |
| SYNCING | Network restored, processing queue | ONLINE, CONFLICT_REVIEW |
| CONFLICT_REVIEW | Sync conflicts detected | SYNCING (resolved), ONLINE (override) |
7.13 Integration Sync States
| State | Description | Transitions To |
|---|---|---|
| IDLE | No sync operation in progress | SYNCING |
| SYNCING | Active sync operation underway | COMPLETED, FAILED, PARTIAL |
| COMPLETED | Sync finished successfully | IDLE |
| FAILED | Sync failed after all retries exhausted | IDLE (manual retry), SYNCING (auto-retry) |
| PARTIAL | Some records synced, others failed | IDLE (accept partial), SYNCING (retry failed) |
7.14 Integration Connection States
| State | Description | Transitions To |
|---|---|---|
| NOT_CONFIGURED | Integration not set up | CONNECTING |
| CONNECTING | Validating credentials and testing connection | CONNECTED, ERROR |
| CONNECTED | Active and operational | DISCONNECTED, ERROR, RATE_LIMITED |
| DISCONNECTED | Manually disabled by admin | CONNECTING |
| ERROR | Connection failed or credentials invalid | CONNECTING (re-validate), NOT_CONFIGURED (reset) |
| RATE_LIMITED | API rate limit exceeded, temporarily paused | CONNECTED (after cooldown) |
7.15 Amazon Order Fulfillment States (FBM)
| State | Description | Transitions To |
|---|---|---|
| PENDING | Order received from Amazon, awaiting payment confirmation | ASSIGNED, CANCELLED |
| ASSIGNED | Routed to a POS store location for fulfillment | PICKING, CANCELLED |
| PICKING | Store staff locating and scanning items | PACKED, CANCELLED |
| PACKED | Items packaged and ready for shipment | SHIPPED |
| SHIPPED | Carrier picked up, tracking number provided to Amazon | DELIVERED, EXCEPTION |
| DELIVERED | Carrier confirms delivery | (terminal) |
| CANCELLED | Order cancelled by customer or seller | (terminal) |
| EXCEPTION | Delivery exception (lost, damaged, refused) | SHIPPED (re-attempt), CANCELLED (refund) |
7.16 Product Sync Validation States
| State | Description | Transitions To |
|---|---|---|
| DRAFT | Product created, not yet validated for any channel | VALIDATING |
| VALIDATING | Running cross-platform validation rules | VALID, INVALID |
| VALID | All required fields pass validation for target platform(s) | SYNCING, INVALID (data changed) |
| INVALID | One or more validation rules failed | VALIDATING (after fix) |
| SYNCING | Pushing product data to external platform | SYNCED, SYNC_FAILED |
| SYNCED | Successfully published to external platform | VALID (re-validate on change), SYNCING (update) |
| SYNC_FAILED | Push to platform failed | SYNCING (retry), VALID (re-queue) |
| BLOCKED | Product excluded from sync (e.g., exceeds variant limits) | VALIDATING (after fix) |
Appendix M: Field Specifications Reference
This appendix summarizes the 44 field-level specifications gathered through requirements interviews. Complete Technical User Stories with field-level acceptance criteria are documented in the companion document: Technical-User-Stories-Field-Specs.md (60 stories, 5,285 lines).
M.1 SKU & Barcode Specifications
| Specification | Value |
|---|---|
| SKU Max Length | 20 characters |
| SKU Allowed Characters | Alphanumeric + dash + underscore ([A-Z0-9\-_]) |
| SKU Uniqueness | Globally unique per tenant |
| Barcode Types Supported | UPC-A (12 digit), EAN-13 (13 digit), Internal |
| Invalid Barcode Behavior | Show error + manual SKU entry option |
M.2 Customer Contact Specifications
| Specification | Value |
|---|---|
| Phone Format | E.164 international (+1-555-123-4567) |
| Required Customer Fields | First name + Last name + (Phone OR Email) |
| Address Format | Structured fields (Street/City/State/ZIP) |
| ZIP Code Validation | 5 digits or 9 digits (12345 or 12345-6789) |
| Customer Notes Max Length | 500 characters |
M.3 Pricing Specifications
| Specification | Value |
|---|---|
| Price Data Type | DECIMAL(10,2) |
| Price Range | $0.00 - $99,999.99 |
| Tax Rate Format | Percentage with 2 decimals (stored as 8.25) |
| Zero Price Handling | Allowed with mandatory reason code (SAMPLE, DONATION, PROMO, OTHER) |
M.4 Inventory Reason Codes
| Category | Values |
|---|---|
| Adjustment | SHRINKAGE, DAMAGE, COUNT_CORRECTION, VENDOR_ERROR, FOUND_STOCK, SAMPLE, DONATION, EMPLOYEE_PURCHASE, OTHER |
| Transfer | REBALANCE, REPLENISHMENT, CUSTOMER_REQUEST, OVERSTOCK, CONSOLIDATION, OTHER |
| Non-PO Receive | SAMPLE, REPLACEMENT, FOUND_STOCK, CONSIGNMENT, DONATION, VENDOR_CREDIT_RETURN, RMA_RETURN, OTHER |
| Custom Codes | Predefined + Admin-created custom codes allowed |
M.5 Authentication Specifications
| Specification | Value |
|---|---|
| POS PIN Format | 4 digits numeric only (regex: ^\d{4}$) |
| Lockout Policy | 5 failed attempts → 15-minute lock, admin can unlock |
| Password Requirements | 8+ chars, 1 uppercase, 1 number |
| Manager Override | Manager must enter their own PIN (creates audit trail) |
M.6 Payment Specifications
| Specification | Value |
|---|---|
| Card Minimum | None (accept any amount) |
| Supported Card Types | Visa, Mastercard, Amex, Discover |
| Payment Decline Message | “Payment declined. Please try another payment method.” |
| Cash Maximum | $10,000 (IRS reporting threshold) |
M.7 Product Variant Specifications
| Specification | Value |
|---|---|
| Size Values | Predefined by category (Apparel: XS-3XL, Shoes: 5-15, Pants: 28-44), admin can add |
| Color Values | Predefined list (20+ standard) + admin custom additions |
| Max Dimension Values | 50 per dimension |
| Dimension Value Max Length | 30 characters |
M.8 Custom Field Specifications
| Specification | Value |
|---|---|
| Dropdown Max Options | 25 options per dropdown |
| Text Field Max Length | 255 characters |
| Number Field Precision | DECIMAL(10,4), max 999,999.9999 |
| Fields Per Entity | 50 max per entity type (Product/Customer/Order/Vendor) |
M.9 Approval Workflow Specifications
| Specification | Value |
|---|---|
| Threshold Type | Configurable per action ($ or %) |
| Timeout Action | Auto-reject after 48 hours |
| Self-Approval | Not allowed — different person must approve |
| Default Thresholds | Fully configurable by Tenant Admin during setup |
M.10 Receipt & Email Specifications
| Specification | Value |
|---|---|
| Receipt Line Length | 40 characters per line (80mm thermal) |
| Email Subject Max Length | 100 characters |
| Email Body Max Size | 50 KB (HTML) |
| Merge Field Syntax | {{FIELD_NAME}} |
M.11 Time & Date Specifications
| Specification | Value |
|---|---|
| Time Format Display | 12-hour AM/PM (e.g., 9:00 AM) |
| Time Storage | 24-hour format |
| Clock-In Duration | Maximum 16 hours; auto-alert to manager if clock-out not recorded within 16 hours |
| Default Timezone | America/New_York (Eastern) |
| Date Format Display | MM/DD/YYYY |
| Date Storage | ISO format (YYYY-MM-DD) |
M.12 Error Message Specifications
| Specification | Value |
|---|---|
| Error Max Length | 80 characters |
| Error Code Format | Prefix with [ERR-XXXX] |
| Error Tone | Direct and instructive (e.g., “Enter a valid email address.”) |
| Error Display | Inline below each field |
M.13 Error Code Ranges
| Module | Error Code Range |
|---|---|
| Module 1: Sales | ERR-1001 to ERR-1099 |
| Module 2: Customers | ERR-2001 to ERR-2099 |
| Module 3: Catalog | ERR-3001 to ERR-3099 |
| Module 4: Inventory | ERR-4001 to ERR-4099 |
| Module 5: Setup | ERR-5001 to ERR-5099 |
| Module 6: Integrations (General) | ERR-6001 to ERR-6009 |
| Module 6: Shopify | ERR-6010 to ERR-6029 |
| Module 6: Amazon SP-API | ERR-6030 to ERR-6049 |
| Module 6: Google Merchant | ERR-6050 to ERR-6069 |
| Module 6: Product Validation | ERR-6070 to ERR-6079 |
| Module 6: Inventory Sync | ERR-6080 to ERR-6089 |
| Module 6: Payment Integration | ERR-6090 to ERR-6094 |
| Module 6: Email Integration | ERR-6095 to ERR-6097 |
| Module 6: Shipping Integration | ERR-6098 to ERR-6099 |
Companion Document: For complete Technical User Stories with field-level acceptance criteria, see:
Technical-User-Stories-Field-Specs.md
Document History
| Version | Date | Changes |
|---|---|---|
| 10.0 | 2026-01-26 | Initial unified specification |
| 11.0 | 2026-01-26 | Added: Gift Cards, Dedicated Exchanges, Price Tiers, Special Orders, Multi-Store Inventory (full payment required), Commissions, Return Policy Engine, Serial Numbers, Hold for Pickup, Cash Drawer Management, Price Check Mode, Coupons, Flexible Loyalty, Customer Groups, Customer Notes, Communication Preferences |
| 11.1 | 2026-01-26 | Added: State Machine Diagrams (8 entities), Gherkin Acceptance Criteria (Sales & Customers), Business Rules Configuration (YAML), State Machine Reference section |
| 12.0 | 2026-01-26 | Major Update: Added Section 1.16 Offline Operations (queue-and-sync), Section 1.17 Tax Calculation Engine (custom, Virginia-first with expansion design), Section 1.18 Payment Integration (SAQ-A semi-integrated). Fixed inconsistencies: Hold for Pickup state machine reconciled, Discount calculation order clarified (added loyalty redemptions), Credit limit calculation documented (includes pending layaways), Void vs Return distinction clarified (same-day vs after). Added missing user stories: Payment failures, Receipt reprinting, Privacy compliance (GDPR-style). Added Parked Sale state machine. Updated Commission rules for proportional reversal on returns. Added Gift Card jurisdiction compliance (California cash-out). Added Customer self-service and privacy workflows. New state machines: Parked Sales (3.2), Offline Mode (3.11). Updated all Gherkin acceptance criteria. |
| 13.0 | 2026-02-01 | Major Update: (1) Renamed all RFID references to Scanner for technology-agnostic input. (2) Removed Cash Rounding feature. (3) Added Affirm third-party financing (BNPL) as payment method. (4) Receipt scanning now required for return validation. (5) Added X-Report for mid-shift cash audits. (6) New Section 1.7.3 Ship to Customer from Other Location with carrier integration. (7) Card refund now offers staff choice between manual on terminal or automatic via token. (8) Multiple credit cards and cash+card(s) combinations supported. (9) Renamed Integrated Card to Credit Card for clarity. (10) Added Exchange to toggle mode (Sale/Return/Exchange). (11) Loyalty Redemption now applies after tax calculation. (12) Return/exchange policy is configurable in Settings/Setup, not hardcoded. (13) Offline mode notifies customers via email when transfer/ship/reserve items sold at source. (14) Added Reports & Email Templates sub-sections under all major sections (50 reports, 6 email templates). (15) Clarified Reservation vs Hold for Pickup distinction with BOPIS examples. (16) Added Online/In-Store return and exchange policy examples. |
| 14.0 | 2026-02-02 | Major Catalog Expansion: Expanded Section 3 (Catalog Module) from 8 subsections to 23 subsections. (1) New Pricing Engine with 5-level price hierarchy, price books, 4 promotion types, markdown workflows with accountability controls. (2) New Multi-Channel Management with visibility controls, inventory allocation, and channel-specific pricing. (3) New Shopify Integration with POS-master sync strategy, field-level ownership model, and optional bi-directional mode. (4) New Vendor RMA workflow (8-state machine). (5) New Reorder Management with velocity-based dynamic reorder points and auto-generated draft POs. (6) New Inventory Control with 6 statuses, 5 counting methods, approval-gated adjustments, and unified receiving workflow. (7) New Inter-Store Transfers with state machine and auto-rebalancing. (8) New Serial & Lot Tracking. (9) New Landed Cost & Weighted Average Costing. (10) New Product Movement History with stock ledger. (11) New Product Search & Discovery with full-text search, filters, substitutions. (12) New Label & Price Tag Printing with templates. (13) New Product Media management (images + video). (14) New Product Notes & Attachments with structured types. (15) New Catalog Permissions & Approvals (RBAC, field-level, approval workflows). (16) New Product Performance Analytics (ABC classification, embedded metrics). (17) Expanded Product Data Model with retail attributes, custom fields, UoM, shipping, templates, matrix management. (18) Expanded Categories to include formal seasons and reporting dimensions. (19) Expanded User Stories with 14 new epics (F through S) and comprehensive Gherkin acceptance criteria. (20) Added “Features Not Needed” section documenting explicit exclusions (warranties, consignment, expiration, assembly, recalls, product-level tax). |
| 15.0 | 2026-02-04 | New Inventory Module: (1) Created Module 4: Inventory Management with 19 sections (4.1-4.19). (2) Moved inventory content (PO, RMA, Transfers, Counts, Costing, Movement History) from Catalog Module 3 to dedicated Inventory Module 4. (3) Reduced Catalog Module from 23 to 15 sections (3.1-3.15). (4) New sections: POS & Sales Integration (reserve/commit model), Online Order Fulfillment (nearest-store), Offline Inventory Operations (queue + conflict resolution), Alerts & Notifications (5 types + 4 email templates), Inventory Dashboard & Reports (dedicated KPIs + 33 reports), Inventory Business Rules YAML. (5) Added 16 user story epics (4.A-4.P) with 42 stories and 10 Gherkin feature files (52 scenarios). (6) Added cross-references between Modules 1, 3, and 4. (7) Renumbered State Machine Reference to Section 5 and Business Rules to Section 6. (8) Added 20 new decisions (#35-54) covering receiving, counting, transfers, POS integration, offline ops, fulfillment, alerts, and dashboard. |
| 16.0 | 2026-02-04 | New Setup & Configuration Module: (1) Created Module 5: Setup & Configuration with 21 sections (5.1-5.21). (2) System settings with core, operational, and branding configuration. (3) Multi-currency support with USD base and manual exchange rates. (4) Flat location hierarchy (Location → Zones) with predefined and custom zones. (5) User profiles with predefined roles (Staff/Manager/Admin/Buyer/Owner) and configurable feature toggles. (6) Register management with device pairing and two profiles (Full POS, Mobile Checkout). (7) Central printer registry with register linking. (8) Simple flat tax rate per location. (9) Predefined + custom UoMs with conversion factors. (10) Payment method configuration per location with processor settings. (11) Per-entity custom fields (Product/Customer/Order/Vendor). (12) Per-action approval workflows with escalation. (13) Full receipt builder with email receipt template. (14) Central email template registry with SMTP configuration. (15) Integration hub for Shopify, payments, and email providers. (16) Loyalty settings split from Module 2 (rules in M2, settings in M5). (17) Configurable audit logging with retention and export. (18) Consolidated all Business Rules YAML from old Section 6 and Section 4.18 into Module 5.19, organized by module. (19) 14-step tenant onboarding wizard with go-live validation. (20) 17 user story epics (5.A-5.Q) with 51 Gherkin scenarios. (21) Removed old Section 6 (Business Rules Configuration). (22) Renumbered State Machine Reference to Section 6. (23) Added 17 new decisions (#55-71). |
| 17.0 | 2026-02-06 | Field Specifications & Technical User Stories: (1) Completed 12-round requirements interview gathering 44 field-level specifications. (2) Created 60 Technical User Stories with field-level acceptance criteria in companion document. (3) Added Appendix M: Field Specifications Reference summarizing all validation rules, data types, formats, and error codes. (4) Established error code ranges: ERR-1xxx (Sales), ERR-2xxx (Customers), ERR-3xxx (Catalog), ERR-4xxx (Inventory), ERR-5xxx (Setup). (5) Documented reason codes for inventory adjustments, transfers, and non-PO receiving. (6) Defined authentication specs (4-digit PIN, 8+ char password, 5 failures = 15-min lockout). (7) Specified product variant constraints (50 values per dimension, 30 chars each). (8) Set approval timeout to auto-reject after 48 hours. (9) Added 12 new decisions (#72-83) for field-level specifications. |
| 18.0 | 2026-02-17 | New Integrations Module & Multi-Platform Expansion: (1) Created Module 6: Integrations & External Systems with 13 sections (6.1-6.13). (2) Moved integration content from Modules 1, 3, 4, and 5 into Module 6 with redirect stubs at original locations. Moved sections: 3.7 Shopify Integration → 6.3, 4.14.3 Inventory Sync with Shopify → 6.3.14, 5.11.3 Payment Processor Configuration → 6.8.3, 5.15.1 Email Provider Configuration → 6.9.1, 5.16 Integrations Hub → 6.11. (3) New Section 6.2: Integration Architecture with provider abstraction, retry/backoff, circuit breaker, idempotency framework, rate limit management, and webhook processing pipeline. (4) Enhanced Shopify Integration (Section 6.3) with GraphQL API preference, @idempotent directive (mandatory 2026-04), Bulk Operations API, POS UI Extensions, 2026 rate limits, webhook topics catalog, third-party POS integration rules, sync rules & best practices (single source of truth, location config, real-time sync, omnichannel/BOPIS, staff security), and hardware compatibility. (5) New Amazon SP-API Integration (Section 6.4) with OAuth via LWA, Catalog Items API, Listings Items API, Orders API (2-min polling), FBA Inventory API (ASIN/SKU/FNSKU mapping, 7 inventory states), SQS/EventBridge notifications, per-endpoint rate limits, and comprehensive compliance & seller requirements (seller code of conduct, packaging/labeling, FBA vs FBM support, safety buffers, order routing). (6) New Google Merchant API Integration (Section 6.5) with service account auth, ProductInput/Product resource split, local inventory sync (storeCode-level), push notifications, rate limits, required product data fields (11 mandatory + 8 recommended), image quality requirements (8 rules), disapproval prevention rules (10 policies with automated pre-sync validation), Google Business Profile integration, and Content API migration plan (EOL 2026-08-18). (7) New Cross-Platform Product Data Requirements (Section 6.6) with unified validation matrix (strictest-rule-wins), image requirements matrix, pre-sync validation engine with data model, and platform-specific product attributes for Amazon, Google, and Shopify. (8) New Cross-Platform Inventory Sync Rules (Section 6.7) with real-time sync architecture, safety buffer configuration (3 calculation modes: FIXED/PERCENTAGE/MIN_RESERVE), oversell prevention (reserve-on-order, first-commit-wins), channel-specific inventory rules, and graduated sync failure handling. (9) Consolidated Payment Processor Integration (Section 6.8), Email Provider Integration (Section 6.9), Carrier & Shipping Integration (Section 6.10), and enhanced Integration Hub (Section 6.11) with AMAZON and GOOGLE_MERCHANT added to integration_type enum. (10) New Integration Business Rules YAML (Section 6.12) covering shopify, amazon, google_merchant, product_validation, inventory_sync, and global settings. (11) New Integration User Stories (Section 6.13) with 7 epics (6.A-6.G) and 10 Gherkin feature files covering Shopify sync, Amazon orders, Google local inventory, cross-platform validation, safety buffers, payment terminals, and integration hub. (12) Renumbered State Machine Reference from Section 6 to Section 7. (13) Added 4 new state machines: Integration Sync States (7.13), Integration Connection States (7.14), Amazon Order Fulfillment States (7.15), Product Sync Validation States (7.16). (14) Added ERR-6001 to ERR-6099 error code range to Appendix M.13 with sub-ranges per provider. (15) Added 16 new decisions (#84-99) covering module structure, API choices, compliance rules, validation strategy, safety buffers, and dual fulfillment support. |
| 19.0 | 2026-02-19 | Tax Redesign & Simplification Update: (1) Replaced flat tax_rate with tax_jurisdiction_id FK supporting 3-level compound taxes (State/County/City). (2) Added is_franchise Boolean to locations. (3) Removed zone tracking from all modules. (4) Removed role-based location access enforcement; user_locations informational only. (5) Simplified shift management to clock-in/clock-out. (6) Added register IP modification limit (2/365 days). (7) Restricted register retirement to OWNER with type-to-confirm. (8) Added Decisions #100-107. |
Decision Log
Decisions captured during BRD review and refinement:
| # | Decision | Choice | Rationale | Date |
|---|---|---|---|---|
| 1 | Offline Strategy | Queue-and-sync | Allows continued operation, sync on reconnect | 2026-01-26 |
| 2 | Tax Engine | Build custom | Full control over jurisdiction rules, expansion flexibility | 2026-01-26 |
| 3 | Payment PCI Scope | SAQ-A | Simplest compliance, card data never touches system | 2026-01-26 |
| 4 | Multi-tenant Isolation | Schema-based | Balance of isolation and operational efficiency | 2026-01-26 |
| 5 | Commission Reversal | Proportional on returns | Fair to employees, full reversal only on voids | 2026-01-26 |
| 6 | Geographic Scope | Virginia → US → International | Design for most restrictive jurisdiction from start | 2026-01-26 |
| 7 | Gift Card Default | No expiry (California rules) | Most restrictive as baseline, enable where permitted | 2026-01-26 |
| 8 | Discount Order | Added loyalty redemptions before tax | Complete calculation order documented | 2026-01-26 |
| 9 | Credit Limit | Include pending layaways | Accurate available credit calculation | 2026-01-26 |
| 10 | Void vs Return | Void = same day only | Clear distinction for commission handling | 2026-01-26 |
| 11 | Scanner Terminology | Replace RFID with Scanner | Technology-agnostic input device naming | 2026-02-01 |
| 12 | Cash Rounding | Removed | Not required for business operations | 2026-02-01 |
| 13 | Third-Party Financing | Affirm as BNPL provider | Customer financing option, store receives full payment immediately | 2026-02-01 |
| 14 | Receipt Validation | Mandatory scanning before returns | Prevents fraudulent returns, system validates receipt authenticity | 2026-02-01 |
| 15 | X-Report | Mid-shift cash audit (does not close drawer) | Enables shift handoffs and spot-checks without closing drawer | 2026-02-01 |
| 16 | Ship to Customer | Direct shipping from source store | Carrier API integration for real-time shipping cost calculation | 2026-02-01 |
| 17 | Card Refund Method | Staff choice: manual or auto via token | Flexibility for customer-present and customer-absent scenarios | 2026-02-01 |
| 18 | Multi-Card Payment | Multiple cards + cash+card(s) allowed | Each card token stored separately for individual refund processing | 2026-02-01 |
| 19 | Loyalty After Tax | Redemption applies after tax calculation | Loyalty discount reduces final total including tax | 2026-02-01 |
| 20 | Return Policy Config | Manually configured in Settings/Setup | Per-tenant, per-store, per-channel (online vs in-store) | 2026-02-01 |
| 21 | Offline Sold Notification | Email customer via TMPL-OFFLINE-SOLD | Customer informed when transfer/ship/reserve item unavailable | 2026-02-01 |
| 22 | Hold for Pickup Scope | In-store holds + BOPIS | Clear distinction from Reservation (different store) | 2026-02-01 |
| 23 | Pricing Model | Centralized + Price Books + Channel overrides | 5-level hierarchy enables flexible pricing without complexity | 2026-02-02 |
| 24 | Shopify Sync | POS-master default, optional bi-directional | Industry standard (Lightspeed, Retail Pro, SKU IQ all use POS-master); bi-directional option for 3rd party Shopify editors | 2026-02-02 |
| 25 | Shopify Field Ownership | Per-field direction model | Eliminates conflicts by assigning clear ownership; SEO stays in Shopify, product data stays in POS | 2026-02-02 |
| 26 | Inventory Sync | Always bidirectional | Inventory quantities sync both ways regardless of catalog sync mode | 2026-02-02 |
| 27 | Consignment | Not supported | All inventory purchased outright; no consignment tracking needed | 2026-02-02 |
| 28 | Warranties | Not supported | Warranty tracking handled outside POS system | 2026-02-02 |
| 29 | Product Expiration | Not applicable | Clothing/accessories business; no expiration dates needed | 2026-02-02 |
| 30 | Product Assembly | Not needed | Bundles are virtual pricing groupings only, not physical assembly | 2026-02-02 |
| 31 | Product Tax | Location-based only | No product-level tax variation; tax determined by store jurisdiction | 2026-02-02 |
| 32 | Reorder Strategy | Velocity-based dynamic | Dynamic reorder points from sales velocity, not static thresholds; seasonal adjustment | 2026-02-02 |
| 33 | Costing Method | Weighted average | Recalculated on every PO receive; used for COGS and margin | 2026-02-02 |
| 34 | Markdown Accountability | Formal workflow + approval | All price changes tracked with who/when/old/new/reason; manager approval required | 2026-02-02 |
| 35 | Receive Mode | Open receive | Staff sees expected qty; faster workflow; variances still recorded | 2026-02-04 |
| 36 | Receiving Discrepancies | Triple approach | Note variance + auto-RMA draft + quarantine damaged goods | 2026-02-04 |
| 37 | Non-PO Receiving | Allow with reason code | Supports samples, replacements, return-to-stock, found stock | 2026-02-04 |
| 38 | Over-shipment | Threshold-based | Allow up to configurable %; above requires manager approval | 2026-02-04 |
| 39 | Adjustment Approval | All require manager | Strongest control; every adjustment must be reviewed and approved | 2026-02-04 |
| 40 | Custom Reason Codes | Standard + tenant-defined | Standard set plus ability to add custom codes per tenant | 2026-02-04 |
| 41 | Count Freeze | Configurable per count | Manager chooses freeze or snapshot per count session | 2026-02-04 |
| 42 | Count Input | Scanner-primary | Barcode scan increments by 1; manual override for damaged barcodes | 2026-02-04 |
| 43 | Transfer Initiation | Both directions + auto-suggest | HQ push + store pull + system auto-suggests from supply imbalances | 2026-02-04 |
| 44 | Allocation Strategy | Manager manual | Manual allocation when multiple stores need scarce item | 2026-02-04 |
| 45 | REMOVED in v19.0 — zones eliminated; inventory tracked per-location only | 2026-02-04 | ||
| 46 | Overstock Returns | Supported | Negotiated return of seasonal/end-of-line unsold goods to vendor | 2026-02-04 |
| 47 | Sale Decrement | Reserve + commit | Reserve on cart add, commit on payment; reserve for parked/held | 2026-02-04 |
| 48 | Offline Inventory | Queue + conflict resolution | Queue all changes offline; conflict resolution on reconnect | 2026-02-04 |
| 49 | Min Display Qty | Advisory only | Soft warning; doesn’t block sales or transfers | 2026-02-04 |
| 50 | Return to Stock | Auto to available | Customer returns auto-go to Available; staff marks damaged separately | 2026-02-04 |
| 51 | Online Fulfillment | Nearest store | Reserve inventory from store closest to customer shipping address | 2026-02-04 |
| 52 | PO Approval | Threshold-based | Auto-approve under configurable $; manager approval above | 2026-02-04 |
| 53 | Inventory Dashboard | Dedicated | Standalone dashboard with inventory KPIs separate from main admin | 2026-02-04 |
| 54 | Dead Stock | Alert + report only | Flag no-sales items; manager decides action manually | 2026-02-04 |
| 55 | Permission Model | Predefined roles + feature toggles | Balance between simplicity and flexibility; roles are fixed, feature access is configurable | 2026-02-04 |
| 56 | Location Hierarchy | Flat (Locations only) | Simple single-level sufficient for current scale; no multi-region complexity | 2026-02-04 |
| 57 | Register Config | Full (list + device + profiles) | Complete hardware management; two profiles (Full POS, Mobile) cover all use cases | 2026-02-04 |
| 58 | Tax Model | Originally flat rate; upgraded to 3-level compound model (State/County/City) per Decision #100 | 2026-02-04 | |
| 59 | UoM Approach | Predefined + custom with conversions | Standard units provided; custom UoMs with conversion factors for flexible product measurement | 2026-02-04 |
| 60 | Supplier Config | Payment terms + lead times only | Lean setup; full vendor data stays in Module 3 Section 3.8 | 2026-02-04 |
| 61 | Custom Fields | Per-entity, no field groups | Simple field management per entity type; no grouping or validation rules needed | 2026-02-04 |
| 62 | Approval Workflows | Per-action rules | Each approvable action configured independently; granular control without matrix complexity | 2026-02-04 |
| 63 | Email Templates | Central registry + SMTP | Single management point for all templates; SMTP/provider config centralized | 2026-02-04 |
| 64 | System Branding | Full suite (core + operational + branding) | Complete tenant customization including login page, receipt branding, report headers | 2026-02-04 |
| 65 | YAML Consolidation | All into Module 5 | Single source of truth; removes Section 6 and absorbs Section 4.18 | 2026-02-04 |
| 66 | Receipt Config | Full builder | Field selection, ordering, sizing, paper width — maximum receipt customization | 2026-02-04 |
| 67 | Payment Config | Methods + processors in Module 5 | Centralizes payment setup with per-location enable/disable and processor credentials | 2026-02-04 |
| 68 | Currency | Multi-currency, USD base, manual rates | Supports vendor POs in vendor currency; manual rate management sufficient | 2026-02-04 |
| 69 | Audit Logging | Configurable categories | Admin toggles which actions are logged; configurable retention and export | 2026-02-04 |
| 70 | Loyalty Settings | Split (rules M2, settings M5) | Business logic stays with customer module; configurable values centralized in Setup | 2026-02-04 |
| 71 | Tenant Onboarding | Step-by-step wizard | Documented 14-step setup flow for new tenant provisioning | 2026-02-04 |
| 72 | SKU Format | 20 chars, alphanumeric + dash/underscore | Industry standard; covers most retail SKU schemes | 2026-02-06 |
| 73 | Barcode Types | UPC-A, EAN-13, Internal only | Standard retail barcodes; internal for custom use | 2026-02-06 |
| 74 | Phone Format | E.164 international | Supports global customers with standardized format | 2026-02-06 |
| 75 | Price Precision | DECIMAL(10,2), max $99,999.99 | Standard retail pricing; sufficient for high-value items | 2026-02-06 |
| 76 | POS PIN Format | 4 digits numeric | Fast entry on touchscreen; familiar to retail staff | 2026-02-06 |
| 77 | Password Rules | 8+ chars, 1 upper, 1 number | Standard strength; balances security with usability | 2026-02-06 |
| 78 | Card Decline Message | Generic “Payment declined” | Protects customer privacy; reduces fraud hints | 2026-02-06 |
| 79 | Variant Dimensions | 50 values max, 30 chars each | Covers extensive size/color ranges with reasonable limits | 2026-02-06 |
| 80 | Custom Field Limits | 50 fields/entity, 25 dropdown options | Generous limits without performance impact | 2026-02-06 |
| 81 | Approval Timeout | Auto-reject after 48 hours | Prevents stuck requests; requester re-submits if needed | 2026-02-06 |
| 82 | Error Code Format | [ERR-XXXX] prefix | Easy support reference; traceable in logs | 2026-02-06 |
| 83 | Error Display | Inline below each field | Immediate visual feedback; industry standard UX | 2026-02-06 |
| 84 | Dedicated Integration Module | Create Module 6 for all integrations | Consolidates scattered integration content from 5 modules into single authoritative source; reduces duplication and cross-reference complexity | 2026-02-17 |
| 85 | Amazon SP-API Authentication | OAuth 2.0 via Login with Amazon (LWA) | Amazon’s required auth method; 1-hour tokens with automatic refresh; regional endpoint support | 2026-02-17 |
| 86 | Google Merchant API Version | Merchant API v1 (migrate from Content API) | Content API for Shopping EOL August 18, 2026; v1 is the successor API with ProductInput/Product resource split | 2026-02-17 |
| 87 | Shopify API Preference | GraphQL over REST | GraphQL has higher rate limits (50pts/sec vs 2req/sec), more efficient queries, better pagination; Shopify’s recommended approach | 2026-02-17 |
| 88 | Shopify Idempotency | @idempotent directive mandatory on all mutations | Required by Shopify starting 2026-04; prevents duplicate operations on retry; uses SHA-256 idempotency key | 2026-02-17 |
| 89 | Amazon Notifications | SQS push notifications | Real-time event delivery for order changes, inventory events, listing status; avoids polling overhead | 2026-02-17 |
| 90 | Google Local Inventory | Store-level local inventory sync | Google requires storeCode-level granularity; maps directly to POS per-location inventory model | 2026-02-17 |
| 91 | Integration Error Codes | ERR-6xxx range | Dedicated error code range for integration module; sub-ranges per provider (6010-6029 Shopify, 6030-6049 Amazon, 6050-6069 Google) | 2026-02-17 |
| 92 | Circuit Breaker | 5 failures in 60 seconds triggers OPEN | Standard resilience pattern; 30-second cooldown before HALF_OPEN probe; prevents cascade failures to external APIs | 2026-02-17 |
| 93 | Provider Abstraction | IIntegrationProvider interface | Common interface for all providers (Connect, Sync, GetStatus, ValidateCredentials); enables consistent error handling and monitoring | 2026-02-17 |
| 94 | Amazon Sync Direction | POS-master default | Consistent with Decision #24 (Shopify POS-master); all external channels receive product data from POS as source of truth | 2026-02-17 |
| 95 | Redirect Stubs | Cross-reference stubs in original locations | Moved sections leave redirect stubs (e.g., “See Module 6, Section 6.3”) to prevent broken references; stubs include brief scope statement | 2026-02-17 |
| 96 | Cross-Platform Validation | Strictest-rule-wins approach | POS enforces the most restrictive requirement across all platforms (e.g., 150-char title from Google, 1000x1000px image from Amazon); ensures products valid everywhere | 2026-02-17 |
| 97 | Safety Buffer | Configurable per-product per-channel inventory buffer | Formula: Channel Available = POS Available - Safety Buffer; prevents overselling during sync delays; critical for Amazon (2-min) and Google (30-min) latency | 2026-02-17 |
| 98 | Dual Fulfillment | Support both FBA and FBM for Amazon | FBA inventory tracked separately (Amazon manages); FBM uses POS pick-pack-ship workflow; per-product fulfillment method configuration | 2026-02-17 |
| 99 | Third-Party POS Rules | Shopify third-party POS integration compliance | Non-native POS must use OAuth, support real-time sync, not bypass checkout; POS is source of truth with field ownership model controlling data flow | 2026-02-17 |
| 100 | Compound Tax Model | 3-level compound (State/County/City) replaces flat rate per location | Virginia model requires state + regional + local stacking; single flat rate insufficient for multi-jurisdiction compliance | 2026-02-19 |
| 101 | Tax Jurisdiction FK | Location references tax_jurisdictions table instead of storing flat rate | Decouples tax configuration from location entity; enables shared jurisdictions and compound rate management | 2026-02-19 |
| 102 | Franchise Flag | is_franchise boolean on locations table | Distinguishes franchise vs company-owned locations for operational rules, reporting, and fee structures | 2026-02-19 |
| 103 | Location Access Informational | user_locations no longer restricts transaction processing; assignments used for defaults and reporting only | Simplifies permission model; all users can operate at any tenant location | 2026-02-19 |
| 104 | Register IP Limit | Max 2 IP address changes per rolling 365 days, tracked in register_ip_changes audit table | Prevents frequent device swapping; ensures hardware stability and audit traceability | 2026-02-19 |
| 105 | Register Retire Safety | OWNER-only with type-to-confirm ‘RETIRE’ | Prevents accidental permanent decommission; strongest safety for irreversible action | 2026-02-19 |
| 106 | Shift Simplification | Simple clock-in/clock-out replaces full shift management (shift types, assignments, handoff notes removed) | Full shift scheduling adds complexity without proportional value; clock-in/out sufficient for time tracking and payroll | 2026-02-19 |
| 107 | Zone Removal | Zones within locations removed; per-location inventory tracking only | Zone sub-divisions added complexity without sufficient operational value; per-location granularity sufficient for current scale | 2026-02-19 |
| 108 | RFID Scope | Counting only — strip lifecycle fields (sold_at, transferred_at) | RFID used exclusively for inventory counting; sales and receiving handled by barcode Scanner; simplifies tag status to (active, void, lost) | 2026-02-25 |
| 109 | Build Custom Raptag | Build .NET MAUI app (not buy off-shelf) | Full control over RFID counting UX, offline-first with SQLite, Zebra SDK integration, multi-operator support | 2026-02-25 |
| 110 | Chunked Upload | 5,000 events per chunk with idempotent dedup | Enterprise scale (100K+ tags) requires chunked sync; UNIQUE(session_id, epc) constraint makes retries safe; resume via upload-status endpoint | 2026-02-25 |
| 111 | Multi-Operator Sessions | Up to 10 operators per session with section assignment | Large stores need parallel counting; session_operators table tracks who scanned where; server deduplicates by highest RSSI | 2026-02-25 |
| 112 | Auto-Save | 30-second SQLite checkpoint flush + session recovery | Protects against data loss from app crashes, battery death; recovery dialog on restart with Resume/Discard options | 2026-02-25 |
| 113 | EPC Serial Numbering | PostgreSQL SEQUENCE per tenant (not column-based) | Concurrent-safe serial assignment; avoids race conditions from last_serial_number column approach | 2026-02-25 |
Document Information
| Attribute | Value |
|---|---|
| Version | 5.0.0 |
| Created | 2026-02-25 |
| Updated | 2026-02-25 |
| Source | BRD v20.0 (19,900+ lines, 7 modules, 113 decisions) |
| Author | Claude Code |
| Status | Active |
| Part | II - Architecture |
| Chapter | 05 of 32 |
Change Log
| Version | Date | Changes |
|---|---|---|
| 4.0.0 | 2026-02-25 | BRD v20.0 integrated as Chapter 08: Architecture Components |
Next Chapter: Chapter 06: Database Strategy
This chapter is part of the POS Blueprint Book. All content is self-contained.