The WMS Integration Playbook: Orders, Inventory and Shipments in Sync
A step-by-step playbook for integrating a warehouse management system with your store, ERP or order system: system-of-record decisions, item and location master data, order release, inventory sync, shipment confirmations, returns, and the reconciliation that keeps stock numbers honest.
- Time needed
- 4 to 10 weeks
- Steps
- 9
- Level
- Advanced
In short: A warehouse management system (WMS) integration connects the system that sells (a store, marketplace or ERP) with the system that picks, packs and ships. Four flows carry almost all the value: item master out, orders out, shipment confirmations back, and inventory back. Most failures are not technical. They come from unclear ownership of stock numbers, mismatched SKUs and units, partial shipments nobody planned for, and events that arrive twice or out of order. This playbook covers the decisions, the flows, the error handling and the daily reconciliation that keeps everyone trusting the numbers.
Key takeaways
- Decide which system owns each piece of data before building anything.
- Get the item master right first: SKU, unit of measure, barcodes, kits.
- Design for partial shipments, cancellations after release, duplicates and out-of-order events.
- Reconcile orders, shipments and inventory every day. Alerts catch errors; reconciliation catches silence.
What you will have live
- A written data ownership map (which system owns items, orders, stock and shipments).
- Item master sync from your product source to the WMS.
- Automatic order release to the WMS with clear rules for holds and cancellations.
- Shipment confirmations with tracking posted back to the selling system.
- Inventory updates from the WMS to every sales channel.
- Returns flow back into stock with a clear disposition.
- Alerts for failures and a daily reconciliation report.
Octacer note: Take one real order with a kit, a backordered line and a split shipment, and trace it on paper through both systems before writing code. That one exercise surfaces most of the mapping questions, and it is far cheaper than finding them on day one of go-live.
Step 1: Decide the system of record
Write down, for each data type, which system is the source of truth and which systems only receive copies.
| Data | Usual owner | Flows to |
|---|---|---|
| Items, SKUs, barcodes, dimensions | ERP or product system | WMS, store |
| Customer orders | Store or ERP | WMS (as fulfillment orders) |
| On-hand stock by location | WMS | ERP, store, marketplaces |
| Available to sell | ERP or order system | Sales channels |
| Shipments and tracking | WMS | Store, ERP, customer notifications |
| Returns received and condition | WMS | ERP, store, refunds |
"On hand" and "available to sell" are different numbers. The WMS knows what is physically on the shelf. Your order system knows what is already promised. Do not let two systems both calculate availability for the same channel.
Step 2: Clean up the item master
The item master is where most integrations quietly break. Before go-live, agree on:
- One SKU per sellable item, identical in every system, or a maintained cross-reference table.
- Units of measure: each, inner pack, case, pallet, with conversion factors.
- Barcodes the warehouse will actually scan.
- Kits and bundles: is the kit a SKU in the WMS, or exploded into components when the order is released?
- Lot, serial and expiry tracking where your products need it.
- Dimensions and weights for packing and carrier rates.
Send item changes to the WMS before any order that uses them. An order line for an unknown SKU should be rejected loudly, not dropped.
Step 3: Release orders to the WMS
Decide when an order becomes a warehouse task. Common rules: payment captured, fraud check passed, address validated, not on hold. Send a stable order ID and line IDs so later messages can refer to them.
Plan for change after release:
- Cancellations before pick are usually simple; after pick or pack they need a process.
- Address changes should be blocked or routed to a person once released.
- Split fulfillment across warehouses means one customer order becomes several warehouse orders.
Shopify's fulfillment service model is a good illustration: a fulfillment service can accept or reject fulfillment requests, merchants can request cancellation before shipment, and one order can contain several fulfillment orders with different delivery methods (Shopify, About fulfillment services). Whatever your platform, model those states explicitly.
If your warehouse is a 3PL that works over EDI, the same flows map to documents such as the 940 shipping order and 945 shipping advice (Stedi, X12 940). Our 3PL EDI integration playbook covers that path.
Step 4: Post shipments back
When the WMS confirms a shipment, post back what actually shipped: order and line IDs, quantities, carrier, service, tracking numbers and ship date. Handle:
- Partial shipments: some lines or quantities ship now, the rest later or never.
- Substitutions, if allowed.
- Multiple packages per order, each with tracking.
- Short picks: the system expected stock that was not there. This is also an inventory signal.
Only mark the order fulfilled in the selling system when the confirmation arrives, not when it was released.
Step 5: Sync inventory
Choose how stock moves from the WMS to channels:
- Event-based: the WMS sends changes as they happen. Fast, but you must handle duplicates and ordering.
- Snapshot: the WMS sends full on-hand by SKU on a schedule. Slower, but self-correcting.
- Both: events for speed, a periodic snapshot to correct drift. This is usually the safest choice.
Apply your availability rules (safety stock, reserved quantities, channel allocation) in one place before pushing numbers to channels. If an inventory app is the source of truth for a store, follow the platform's model for inventory locations and levels (Shopify, Inventory management apps).
Step 6: Handle duplicates, ordering and retries
Integrations receive the same message twice and in the wrong order. Shopify's webhook documentation, for example, says it does not guarantee ordering within a topic or across topics for the same resource, and recommends verifying signatures and ignoring duplicate deliveries using the webhook ID header (Shopify, Webhooks). Design for that everywhere:
- Store a unique message or event ID and skip ones you have already processed.
- Use timestamps or version numbers so an old update does not overwrite a newer one.
- Make writes idempotent so a retry cannot double a shipment or stock change. Microsoft's retry pattern guidance warns that retrying a non-idempotent operation can execute it more than once (Microsoft, Retry pattern).
- Put failed messages in a queue a person reviews, with the reason and a replay button.
Step 7: Returns
Define the return flow end to end: return authorization in the selling system, receipt in the WMS, inspection and disposition (restock, refurbish, scrap), then stock and refund updates. Only restock what passed inspection. Send the disposition back so finance and customer service see the same result.
Step 8: Test with real scenarios
Test in a sandbox or test warehouse with scenarios that match your real business:
- Single line, multi-line, kit and backordered orders.
- Cancel before pick and after pick.
- Partial and split shipments, multiple packages.
- Unknown SKU, wrong unit of measure.
- Duplicate and out-of-order messages.
- Returns with each disposition.
- A cycle count adjustment.
Run a short parallel period where possible, comparing what the integration did with what the warehouse actually did.
Step 9: Monitor and reconcile daily
Alerts catch loud failures. Reconciliation catches quiet ones. Every day compare:
- Orders released versus orders acknowledged by the WMS.
- Orders released more than your agreed time ago without a shipment.
- Shipments in the WMS versus fulfilled orders in the selling system.
- On-hand by SKU in the WMS versus what channels show, allowing for reservations.
Investigate every mismatch and fix the cause, not just the number. Our silent integration failure triage playbook covers the diagnosis.
WMS integration checklist
- [ ] Data ownership map signed off
- [ ] SKU, unit of measure, barcode and kit rules agreed
- [ ] Item changes sent before orders that use them
- [ ] Order release rules, holds and cancellations defined
- [ ] Shipment confirmations handle partial, split and multi-package
- [ ] Inventory sync with events plus periodic snapshot
- [ ] Duplicate and out-of-order handling, idempotent writes
- [ ] Failed-message queue with replay
- [ ] Returns with disposition
- [ ] Scenario tests passed
- [ ] Daily reconciliation report with an owner
Common mistakes
- Two systems calculating availability. Channels oversell or undersell depending on which one updated last.
- SKU mismatches fixed by hand. Manual fixes do not scale and are forgotten.
- Marking orders shipped on release. Customers get tracking emails for orders still on the shelf.
- No plan for partial shipments. The first short pick breaks the integration.
- Trusting events alone. Without snapshots or reconciliation, drift accumulates silently.
Frequently asked questions
What is a WMS integration?
A connection between a warehouse management system and the systems that sell or plan, such as an online store, marketplace or ERP, that exchanges items, orders, shipments, inventory and returns automatically.
Should the WMS or the ERP own inventory?
Usually the WMS owns physical on-hand by location, and the ERP or order system owns available to sell after reservations. The important part is that only one system owns each number.
API, EDI or file transfer?
Use what the WMS or 3PL supports well. Modern WMS platforms tend to offer APIs and webhooks; many 3PLs use EDI; some use files over SFTP. The same flows and reconciliation apply to all three.
How often should inventory sync?
Event-based updates as stock changes, plus a full snapshot on a schedule (often nightly or more frequently for fast-moving SKUs) to correct drift.
How long does a WMS integration take?
It depends on the number of flows, channels and how clean the item master is. Plan several weeks for a single warehouse and channel, including testing with the warehouse team.
Who should own it after go-live?
Name an operations owner for reconciliation and exceptions, and a technical owner for the integration itself. Both should see the same failure queue.
How Octacer helps
Octacer has built integrations between stores, ERPs, CRMs and warehouses since 2017. We design the ownership map, flows, error handling and reconciliation so stock numbers stay trustworthy. See logistics and supply chain, CRM and ERP integration, our NetSuite services and integration architecture.
Next step
Write the data ownership table for your business and trace one difficult order through it. If you want help turning that into a working integration, book a call.
Sources
- Shopify, About fulfillment services: https://shopify.dev/docs/apps/build/orders-fulfillment/fulfillment-service-apps
- Shopify, Inventory management apps: https://shopify.dev/docs/apps/build/orders-fulfillment/inventory-management-apps
- Shopify, About webhooks: https://shopify.dev/docs/apps/build/webhooks
- Stedi X12 reference, 940 Warehouse Shipping Order: https://www.stedi.com/edi/x12/transaction-set/940
- Stedi X12 reference, 846 Inventory Inquiry/Advice: https://www.stedi.com/edi/x12/transaction-set/846
- Microsoft, Retry pattern: https://learn.microsoft.com/en-us/azure/architecture/patterns/retry
Keep going
More playbooks
Want it running on your stack?
We implement this playbook for clients, tailored to your stack, your data, and your team's workflow, and handed over documented.