QAD and Walmart Marketplace integration

What moves between them

The main flow runs from QAD into Walmart Marketplace for catalog data and from Walmart back into QAD for orders and money. ml-connector pushes QAD item listings, inventory quantities, and prices into Walmart through the MP_ITEM, inventory, and price feeds, so the storefront reflects the ERP. New Walmart orders are pulled into QAD as sales orders for fulfillment, and once QAD ships, tracking and carrier details are written back to Walmart to acknowledge and confirm the shipment. Walmart reconciliation reports, which are read-only, are pulled into QAD finance so payouts, fees, and refunds can be matched. GL accounts, suppliers, and employees are not part of Walmart Marketplace, so nothing financial is written into Walmart.

How ml-connector handles it

ml-connector stores both credential sets encrypted. On the Walmart side it exchanges the Client ID and Secret for an access token, attaches it as WM_SEC.ACCESS_TOKEN with the required WM_SVC.NAME header and a fresh correlation id on every call, and refreshes the token before the 15-minute expiry rather than waiting for a 401. On the QAD side it accepts the full tenant URL per customer, since QAD publishes no shared base address, and validates entity paths against that instance. Items are keyed by SKU on both sides, so each QAD item maps to one Walmart listing and the matching inventory and order lines. Because Walmart throttles single-record calls and tells sellers to use feeds, ml-connector batches item, inventory, and price changes into feeds, then polls the feed status endpoint until processing completes. Since QAD cloud is pull-only, it polls QAD on a schedule, and it can subscribe to Walmart order and report webhooks where enabled, falling back to polling Walmart orders inside the 180-day window. Order acknowledge and ship calls ride Walmart's state machine, which makes a repeated acknowledge a safe no-op. Walmart returns HTTP 429 when a rate bucket empties, so ml-connector backs off with jitter and retries, and every record carries a full audit trail and can be replayed if a downstream call fails.

A real-world example

A mid-sized consumer products manufacturer runs QAD Adaptive ERP for production, inventory, and finance, and sells a few hundred SKUs on Walmart Marketplace. Before the integration, a small operations team kept Walmart listings, stock, and prices current by exporting spreadsheets from QAD and uploading them by hand, which meant the storefront often showed stale quantities and oversold items that were already committed in the ERP. New Walmart orders were re-keyed into QAD for picking, and finance reconciled Walmart payouts against the ledger manually at month-end. With QAD and Walmart Marketplace connected, inventory and prices flow into Walmart through feeds on a schedule, orders land in QAD automatically with tracking written back, and settlement reports come into finance for matching, so overselling drops and the manual re-keying is gone.

What you can do

• •Push QAD item listings, inventory levels, and prices into Walmart Marketplace through its bulk feed endpoints.
• •Pull new Walmart orders into QAD for fulfillment and write tracking and carrier details back on shipment.
• •Bring Walmart reconciliation reports into QAD finance so payouts, fees, and refunds can be matched.
• •Bridge QAD tenant login and Walmart's OAuth2 token, refreshing the 15-minute token before it expires.
• •Map items by SKU across both systems, with throttle-aware feeds, retries, and a full audit trail on every record.

Questions

Which direction does data move between QAD and Walmart Marketplace?

Catalog data moves from QAD into Walmart, and orders and money move from Walmart back into QAD. ml-connector pushes QAD items, inventory, and prices into Walmart through feeds, pulls Walmart orders into QAD for fulfillment, and writes shipment tracking back to Walmart. Walmart settlement reports are read-only and are pulled into QAD finance, so nothing financial is written into Walmart.

Why does ml-connector use Walmart feeds instead of single-item updates?

Walmart aggressively throttles its single-record sync APIs and states they are intended for emergency use only, capping updates such as single price changes at 100 per hour. For any volume of items, inventory, or prices, ml-connector batches changes into Walmart's async feed endpoints, then polls the feed status until processing finishes. This keeps large catalog syncs from hitting rate limits.

How does the integration handle QAD's tenant-specific URL and Walmart's short-lived token?

ml-connector accepts the full QAD instance URL per customer, since QAD publishes no shared base address, and polls QAD on a schedule because the cloud product has no webhooks. On the Walmart side it exchanges the Client ID and Secret for an access token and refreshes it before the 15-minute expiry, attaching the WM_SEC.ACCESS_TOKEN and WM_SVC.NAME headers on every call, with both credential sets stored encrypted.