WooCommerce B2B: ordering workflows and ERP stock checks

Define B2B ordering requirements, then compare store and ERP inventory using the right references, reservations and reporting periods.

Two women looking at a clipboard beside storage shelves, one wearing a safety vest.

Article content

To reconcile WooCommerce and ERP inventory, compare the same item, unit and warehouse scope for a comparable period. Then establish what each quantity already includes: allocations, committed orders and a safety buffer. A difference between two screens is not a reason to overwrite one system automatically with the other.

In a B2B store, this check also depends on the ordering workflow: account pricing, quotes, purchase orders and payment terms. This guide explains how to define those requirements, then build a usable inventory reconciliation record. The example is fictional; it does not report results from a client store.

Start with the B2B ordering rules

A distributor needs more than a visible catalogue. Write down the decisions the buying process must support before selecting extensions or custom development:

  • Accounts and pricing: who can view each catalogue, obtain negotiated prices and order for a business? Also test an unauthorised account and a signed-out user.
  • Ordering: must buyers enter a purchase order number, buy by the case or meet a minimum? Define the selling unit and quantity rounding rules.
  • Quotes and payment: when does a quote become an accepted order? Who approves terms, and at which step is stock committed?
  • Service: who receives exceptions, can amend an order and confirms a delivery date?

These are requirements, not features guaranteed by every WooCommerce installation. Have your exact combination of extensions and connector demonstrate the workflow. An existing connector may be suitable when its rules are verifiable. Custom work should address a specific unmet need rather than be the automatic choice.

What you need before comparing inventory

Obtain authorised read-only exports of the relevant products or variations, the mapping to ERP items, reservation rules and recent synchronisation logs. Record the versions of WooCommerce, the connector and extensions affecting stock. Limit data to the references, quantities, states and timestamps needed; an inventory check does not require an export of customer contact details.

Begin with a few representative items in a test environment or frozen exports. Involve a business owner who can explain the quantities and someone familiar with the integration. A spreadsheet can support an initial review. Hosting, extensions, API access, the connector, data cleanup, testing and support may carry costs. Estimate that scope instead of adopting a generic hourly rate or project timeline.

If you are investigating an incident, start with connector logs, errors, order notes and event times. Without that evidence, a synchronisation delay remains a hypothesis.

Define the fields and who holds the stock

The WooCommerce REST API v3 product reference distinguishes manage_stock, stock_quantity, stock_status and backorders. Record them separately: “in stock” is not a count, and allowing backorders is a policy distinct from a quantity.

Create a record for each item containing its ERP identifier, WooCommerce product and variation identifiers, SKU, unit, warehouse scope, stock-owning level, quantity meaning, reporting period and backorder rule. Include the evidence and decision owner. Do not compare cases with individual units or all warehouses with a single location.

Variations can have their own stock or draw from a shared parent. The variations API includes parent as a possible manage_stock value; the variable product guide explains inventory modes. A matching-looking SKU is therefore insufficient evidence of a mapping. Confirm identifiers and avoid counting a shared pool several times.

Align the periods behind the numbers

Exports taken thirty minutes apart may span sales or receipts. Record extraction start and end times, the time zone, the period covered and which events have been processed. A paginated export of an active catalogue is not necessarily a single snapshot.

The product API’s date_modified_gmt field describes when the product was modified. By itself, it does not establish that both systems have processed the same stock movements. If you cannot reconstruct a comparable period, classify the item as “not comparable” and obtain another reading. Do not turn uncertainty into a confirmed inventory discrepancy.

Choose a reproducible method for the pilot: frozen data, a quiet period with recorded movements, or consistent exports supplied by both systems. An acceptable freshness limit must come from your operations. The fictional times below do not define a synchronisation standard.

Deduct reservations only once

Ask what “available” means in your ERP and what the connector sends to WooCommerce. A quantity may already be net of certain orders. Before calculating anything, establish which allocations have reduced each value and which holds remain temporary.

WooCommerce distinguishes Pending payment, On hold and Processing. Its order status documentation describes reduced stock for the latter two. The Hold Stock setting applies to pending-payment orders, not those marked On hold. Also check extension rules and order notes; similar labels do not make the states interchangeable.

Illustrative example: a pilot’s agreed rule starts with 12 physical units, subtracts 3 reserved units and a 1-unit safety buffer, leaving 8 to send to the store. The three reserved units are counted once, no additional temporary hold exists in this dataset, and the store value is defined as already net of those deductions. If it is 8, do not subtract 3 again. This is the example’s formula: the field name stock_quantity alone does not establish that meaning in your installation.

Classify results before making corrections

Calculate a difference only after confirming the item, unit, location, period and meaning of the values. In the example, store quantity 10 minus expected quantity 8 gives a discrepancy of +2. That establishes a difference within the agreed scope, not its cause or which system should be overwritten.

  • Match: comparable quantities agree. An explicit zero is a value that can be checked.
  • Quantity discrepancy: the same references and rules produce different numbers; find the movement or rule explaining the difference.
  • Not comparable: a missing quantity, wrong variation, incompatible period or unconfirmed scope. Never substitute zero for an absent value.
  • Policy review: quantities can agree while the store allows backorders contrary to the intended rule.

For each exception, retain initial values, evidence, the confirmed or still-hypothetical cause and proposed action. Before an actual correction, confirm the authoritative source, movements since the reading and consequences for orders. Correct a small authorised scope, read the result in both systems and check that the next synchronisation does not reintroduce the difference. A bulk refresh from an old file can overwrite recent movements.

What our local demonstration checked

On October 28, 2026, a Python 3.9.6 test compared ten entirely synthetic cases under the rule above. All ten expected results were verified: two matches, one numerical discrepancy, six non-comparable cases and one policy review. On November 10, 2026, a new check of the same ten cases, using a script reconstructed from the documented rules, confirmed these outcomes.

  • Expected quantity 8 matches the store; 10 produces a +2 discrepancy.
  • An absent quantity is rejected as non-comparable; an explicit zero on each side matches.
  • A thirty-minute gap, the wrong variation despite identical SKU text and the wrong warehouse scope are flagged.
  • Stock shared at the parent and an unconfirmed quantity meaning remain outside the comparison.
  • Backorders enabled against a “no” rule trigger review even when quantities agree.

References, locations and times are invented. This dataset’s snapshot_at and location_scope metadata describe fictional exports; they are not claimed to be native product API fields. The matching case represents the same instant in UTC and with a time-zone offset. No actual WooCommerce, ERP, connector or API call was run. The test does not validate concurrent orders, real reservations, cancellations, returns, performance or automatic correction.

Make reconciliation part of the store project

Define who owns each type of data: catalogue and pricing, inventory, accepted orders, payment and account follow-up. The ERP and CRM may serve different purposes. Document mappings, errors and retries; a replayed order should be recognisable without creating a second business record. Our ERP and CRM API integration guide develops that approach.

For workflows offered in French and English, check descriptions, units, out-of-stock messages, accounts, quotes and emails in both languages. Keep the same business identifiers behind translated labels. A correct quantity does not make an ambiguous unit or another account’s price acceptable.

Before launch, run acceptance tests in a separate environment: account permissions and pricing, purchase order entry, minimums and pack multiples, quote conversion, state changes, exhausted stock and synchronisation exceptions. Prepare rollback and a way to handle orders created in the meantime. Ask for test evidence, a recovery procedure and a named support owner. Do not infer a project duration or perfect compatibility from a list of logos. The user acceptance testing grid helps formalise these criteria.

Inventory reconciliation checklist

  • The B2B workflow and the point where stock is committed are defined.
  • Item, variation, unit, warehouse scope and stock owner are confirmed.
  • Readings cover a comparable period and movements are traceable.
  • Reservations are deducted once; absent quantities remain unknown.
  • Quantity, status and backorder policy are reviewed separately.
  • Every discrepancy has evidence, an owner and verification after an authorised correction.

If these rules need to connect your store with existing systems, our custom applications and integrations service in Quebec and across Canada can help scope the pilot and its acceptance criteria.

Technical documentation rechecked on November 10, 2026. Initial demonstration run on October 28; new check of the same cases on November 10, 2026.

Explore more articles

Let’s discuss your project.

Tell us what you would like to improve. We’ll help you decide on the next steps.

Start the conversation