Inventory software requirements should describe how physical and commercial events change quantity, state, location, ownership, value and availability. “Real-time inventory” is not testable until the event, timing and decision are defined.
Start with the stock lifecycle and difficult exceptions. Connect each requirement to authoritative master data, integration behavior, reconciliation and an acceptance scenario.
Set the inventory boundary
List entities, locations, channels, item types, ownership models, lots, serials, expiry, units and currencies. Define which planning, purchasing, order, warehouse and finance processes are in scope.
Separate current mandatory complexity from future growth with named triggers.
Model stock events
Use receipt, put-away, reservation, allocation, issue, transfer, count, adjustment, return, quarantine and disposal. For each, define trigger, roles, identifiers, states and downstream effects.
Include cancellation, duplicate and partial completion. Event reversal should not rely on deleting history.
Create an event card
| Field | Question |
|---|---|
| Starting state | What quantity, location and status exist? |
| Event | Which physical or commercial fact occurred? |
| Authority | Who may record or approve it? |
| Result | How do on-hand, available and value change? |
| Exception | What if item, quantity or system is wrong? |
| Evidence | Which record and reconciliation prove it? |
Define inventory states
Distinguish on-hand, available, reserved, allocated, in-transit, quarantined, damaged, consigned and expired as needed. State when each can promise, pick, value or transfer.
Use plain definitions across sales, warehouse, planning and finance. One field labelled status may not carry all dimensions.
Control item master data
Specify item identifier, description, unit, conversion, barcode, pack, location, lot/serial, shelf life, cost and supplier references. Name authority and effective change.
Test unit conversion, duplicate item, replacement and inactive item with open stock.
Write planning scenarios
Define demand sources, lead time, safety stock, reorder, forecast, minimums, constraints and approval. Use ordinary and volatile items. State which recommendation a planner accepts or overrides and why.
Do not require advanced optimisation without sufficient data, owner and decision cadence.
Write fulfilment scenarios
Test normal, partial, backorder, substitute, cancelled, multi-location and priority orders. Trace reservation to shipment and customer promise.
Include concurrency: two channels requesting the same last item. Define the authority and response.
Define counts and adjustments
Specify cycle selection, blind count, recount, variance approval, reason and financial posting. Separate record correction from investigation.
Track repeated variance by item, location and event to improve the source process.
Specify interfaces
For purchase, order, warehouse, commerce and finance events, define identifier, direction, timing, replay, failure and reconciliation. Keep one state authority per layer.
Test delayed, duplicate and out-of-order messages. A successful connection can still create wrong availability.
Set non-functional context
State peak transactions, locations, devices, response needed, offline/degraded operation, recovery, security, audit and retention. Connect each to an operating event.
Test at realistic volume with labels, scanners and network where relevant.
Define costing and valuation boundaries
State which system calculates standard, average, FIFO or other approved cost; when it changes; and how adjustments, landed cost and write-offs post. Connect operational quantity with finance reconciliation.
Test backdated receipt, return, negative inventory and period close according to the selected accounting design.
Model multi-location transfers
Define dispatch, in-transit, receipt, partial, loss, cancellation and ownership. Decide when source availability falls and destination availability rises. Use one transfer identifier.
Test delayed receipt and duplicate confirmation. Stale transit should appear as an owned exception, not permanent invisible stock.
Specify traceability
For lot, serial, expiry or regulated items, state creation, capture, split, merge, recall, quarantine and history. Test forward and backward trace from supplier to customer.
Limit manual override and retain reason where identity affects safety or obligation.
Plan migration and opening stock
Inventory source items, quantities, locations, states, costs, lots, reservations and open orders. Clean identifiers and map units before load. Freeze or control movements during final reconciliation.
Compare counts and values by item/location and sample physical stock. Opening balance is an operational event, not a spreadsheet import alone.
Design operating ownership
Name item-data, planning, purchasing, warehouse, integration and finance owners. Define exception queues and daily/periodic controls. Requirements are incomplete when no team can maintain state and rules.
Include support for device, label, interface and count failures and a degraded procedure.
Run a full scenario pack
Use receipt-to-sale, last-item concurrency, transfer, count variance, return, damaged lot and interface outage. Reconcile quantity, availability, value and evidence after each.
Capture vendor status as standard, configured, integrated, custom or unsupported and retain edition assumptions.
Prioritise and accept
Make a requirement mandatory when failure blocks stock truth, fulfilment, control or acceptable service. Use event scenarios for vendor demonstrations and release tests.
Retain the inventory state dictionary, event cards, master authority, interfaces, controls, gaps and owners as the implementation contract.
After selection, reuse the same pack for configuration and operational rehearsal. Mark every requirement as standard, configured, integrated, procedural or unsupported and attach the acceptance result. This prevents broad demo claims from becoming assumed production behavior and gives future change teams a stable explanation of the inventory design.
Review it after every major stock-process change.