Shopify integration for catalogue and orders in 2026
Connect Shopify to your ERP or other store systems. Map catalogue and order data, follow a NetSuite example and check sync failures before expanding.
- Published
- Last updated
Yes, a Shopify integration can connect catalogue, inventory and order records to an enterprise resource planning (ERP) system or other store software. Choose a connector with the required data flows, or build the exchange through Shopify’s administrative application programming interface (API). Give each field one system that controls its value before switching on updates.
A connected app can still send the wrong quantity or overwrite your product descriptions. Start with one item and one stock location, then follow an order through to fulfilment before widening the connection.
Choose a Shopify integration route
Choose the route by the records you need to move and who will maintain the connection. Shopify supports administrative access to products, customers and orders, including inventory. Each app or connector implements its own subset of that access.
| Route | Concrete task and direction | Access needed | Costs and ongoing responsibility |
|---|---|---|---|
| Native Shopify functions | Import product records into Shopify or export them to a file | Staff permissions for the relevant records | Store subscription and staff time. Your team prepares files and checks imports |
| Marketplace app | Move the objects named by that app between Shopify and its destination | App installation and approval of its requested scopes | App subscription or usage charges where applicable. The vendor maintains its supported flows |
| ERP provider’s connector | Send Shopify orders to the ERP and return supported stock or fulfilment updates | Authorisation for both accounts and the connector’s required roles | Connector entitlement plus setup and mapping work. Provider support covers its own connector |
| Middleware | Transform records between Shopify and several systems | Authorised connections to every participating system | Platform subscription or usage model plus workflow maintenance by your integration team |
| Custom API work | Implement mappings or order handling that packaged connectors don’t cover | A Shopify app with appropriate administrative scopes and destination access | Development, hosting and ongoing API-version support belong to your team |
| File-only transfer | Load a catalogue snapshot from another system into Shopify | Export access at the source and import access at the destination | File preparation and repeated manual processing. It creates no automatic return flow |
These are route categories, not interchangeable promises of coverage. A marketplace listing is a distribution route. An ERP provider or middleware vendor can distribute its connector there too.
Shopify’s catalogue file guide distinguish product files from inventory files for multiple locations. A comma-separated values (CSV) import is useful for a controlled initial load. It does not keep later orders or stock changes in sync.
For any app, record supported objects and directions before installation. Include whether updates create records, change them or delete them. Product support alone says nothing about refunds, customer changes or partial fulfilments.
Assign authority before mapping fields
Assign one writing authority to each field, then document what the receiving system must preserve. An ERP can control stock while Shopify retains customer-facing descriptions. Two systems writing the same field can repeatedly undo each other’s changes.
The following is an illustrative authority map for a retailer that manages physical stock in its ERP.
| Record or field | Writing authority | Destination and rule | Identity to retain |
|---|---|---|---|
| Product and variant identity | Catalogue operations team | Match existing Shopify products and variants to ERP items before creating anything | Shopify product ID, variant ID and external item ID |
| Merchandising description | Shopify content team | Keep Shopify copy when receiving inventory or price updates | Product ID plus any variant-specific detail field |
| Selling price | Shopify pricing team in this example | Preserve the storefront price. Pass the actual order price to the ERP | Variant ID and order line ID |
| Location inventory | ERP inventory team | Publish the mapped sellable quantity to the correct Shopify location | Inventory-item ID, Shopify location ID and ERP location ID |
| Customer details | Shopify for the placed order | Send the customer and address data required to process that order | Shopify customer ID where present and external customer ID |
| Order and order lines | Shopify at checkout | Create the external order once and retain its original reference | Store identity, Shopify order ID and line IDs |
| Fulfilment | ERP fulfilment team | Return supported shipped quantities and tracking to Shopify | External fulfilment reference and matched order lines |
Shopify’s inventory object relationships link each variant to an inventory item. Inventory levels connect that item to individual locations. A product ID alone cannot identify which size has stock in which warehouse.
Use the stock-keeping unit (SKU) as a matching field alongside stable system IDs. A SKU rename must update the match without creating another item. Hold unmatched or duplicate SKUs for review instead of guessing from a title.
Keep the quantity meaning in the map too. Physical stock, committed stock and sellable stock answer different questions. If your source already subtracts allocations, subtracting them again in the connector understates availability.
Plan the app or API connection
Select the store and the first data flows before authorising the connection. Then approve only the access those flows require and keep the initial import narrow.
- Record the exact Shopify store and external account, including their test or production environment. Decide whether the first run updates existing items or creates new ones.
- Define a small batch and an order-import starting boundary. Exclude older orders already recorded in the destination.
- Install or authorise the chosen app. Match its requested permissions to the planned objects and directions before approving them.
- Configure item and location mappings. Set protected storefront fields explicitly so an initial product load cannot overwrite them.
- Run the controlled batch. Compare the destination records and errors with the source before enabling scheduled or event-driven updates.
Shopify’s API guide identifies the GraphQL Admin API as the route for administrative store-data work. The Storefront API supports buyer-facing browsing and carts. A storefront token is not the credential for an ERP to administer inventory or import orders.
For custom development, create a Shopify app and keep its administrative credentials on the integration service. New apps use GraphQL Admin API. Shopify describes REST Admin API as legacy.
Access scopes govern which records an app can read or write. Older-order access needs the relevant additional permission beyond the default 60-day window. Customer-data restrictions and fulfilment-order permissions also affect what a connector can retrieve or change.
A webhook tells the integration that a change occurred. A scheduled job reads or sends records at intervals. Either route still needs mappings and a way to confirm the destination accepted each update.
Record the supported API version and its upgrade owner. Shopify’s API rate limits include calculated query costs for GraphQL Admin API, with capacity depending on the plan. Limit the initial batch and leave capacity for ongoing orders instead of treating a throttle response as a missing record.
Work through a Shopify and NetSuite integration
Use Oracle’s NetSuite Connector when you need its documented Shopify order, product and fulfilment flows. This example concerns Oracle’s connector. Other vendors’ Shopify and NetSuite integrations have their own mappings and settings.
You need Shopify and NetSuite accounts, a NetSuite Connector account and the Shopify connector added to it. Oracle’s connector prerequisites require matching item SKUs and a NetSuite setup that can process orders and fulfilments. Install the Connector SuiteApp with the required integration features enabled. Use a Shopify development store and NetSuite sandbox for the first tests.
Use one Shopify store per connector account. Oracle’s Shopify FAQ says each additional store needs a separate connector account, including stores managed through Shopify Plus. Keep the example to ordinary retail orders. Business-to-business (B2B) flows have additional configuration.
Authorise the accounts and restrict the first flow
- Open NetSuite Connector and select the Shopify connector and intended account. Under Settings > Credentials, enter the shop-name portion of its myshopify address. Choose Save Settings, then Authorize Account and complete the Shopify authorisation. Oracle documents these Shopify connection steps.
- Prepare NetSuite token-based authentication. Oracle requires the enabled NetSuite Connector integration record and its NetSuite Connector Web Services role. Its instructions assign that role to an employee who also has Administrator access and require the relevant subsidiary visibility.
- Under NetSuite > Settings > Credentials, enter the NetSuite account ID, token ID and token secret. Choose Save and Test Connection. Keep the token secret in authorised credential storage. Follow Oracle’s authentication procedure for the full role and token setup.
- Confirm the NetSuite order-import status and the subsidiary receiving this store’s orders. Map the applicable shipping and payment methods. A successful credential test does not establish that an order can be posted.
- Match one existing Shopify variant SKU to its NetSuite item. Oracle’s Price/Quantity Only sync requires an existing storefront item. Review its price mapping so this example retains Shopify’s selling price and merchandising copy.
- Under Shopify > Mappings > Products, add the inventory_quantity mapping. Open its special location mapping and pair the numeric Shopify location ID with the NetSuite location internal ID. Oracle’s multilocation procedure describes these fields.
- Enable inventory tracking for the test item. Limit the first run to that item, order import and the supported fulfilment return flow. Keep historical imports and full catalogue updates off until reconciliation passes.
Oracle’s documentation uses both the FarApp portal name and NetSuite Connector. Its B2B instructions use the NetSuite Connector portal, while some Shopify procedures retain the FarApp address. Select the connector and account inside your authorised NetSuite Connector workspace. The documented Settings > Credentials steps identify the connection being changed.
Full Product sync can include description and title mappings. Oracle’s default mapping connects body_html to salesDescription. Preserve Shopify merchandising copy by excluding that overwrite from the chosen flow before the first run.
Follow one item through an order
This illustrative acceptance example uses invented record labels and quantities. Assume one tracked location with stock sold only through this example. The sellable quantity has no safety stock deduction, and backorders are disabled.
Start with variant SKU MUG-NAVY linked to NetSuite item NS-501 and the recorded Shopify variant and inventory-item IDs. Pair NetSuite location WH-1 with Shopify location SH-1. These location labels stand for the actual IDs in your mapping.
| Action | Expected NetSuite record | Expected Shopify record |
|---|---|---|
| Increase sellable stock from 8 to 12 | NS-501 at WH-1 has 12 available | The matched inventory item at SH-1 has 12 available after inventory sync |
| Buy 2 units through Shopify | One imported order references the Shopify order ID, MUG-NAVY and quantity 2 | One order has that variant and quantity 2. Available stock falls to 10 |
| Allocate the imported order | The order reserves 2 units. The source quantity available for sale is 10 | The next inventory refresh leaves availability at 10 |
| Fulfil 1 of the 2 units | An item fulfilment records quantity 1 against the original order line | That line shows 1 fulfilled and 1 remaining, with the matching tracking information |
Match order totals and shipping charges as well as item quantities. Retain the Shopify order ID in the external order reference, then record the NetSuite internal order ID in the reconciliation log.
Oracle’s fulfilment mapping guidance supports partial or complete fulfilment after a NetSuite item fulfilment is created. Standard shipping information includes the carrier and tracking number. Keep the ordered items unchanged while testing partial fulfilment.
Treat timing and cancellations as separate settings
Scheduled inventory updates follow the product-sync schedule. Separately, Oracle provides an Add Real-Time Sync Webhook option under Shopify > Settings > Orders > Advanced Options. Its real-time order instructions warn that enabling this can increase the subscription charge. That setting does not establish real-time inventory or fulfilment timing.
Cancellation needs its own decision. Oracle’s cancellation settings can send qualifying NetSuite cancellations to Shopify. They include pending-approval orders cancelled in NetSuite and pending-fulfilment orders closed there.
Oracle says paid orders cancelled through this flow refund automatically by default. Its Prevent Shopify From Refunding Paid Orders When Cancelling option changes that behaviour. Decide who issues the refund before enabling cancellation sync, so staff don’t refund the same payment again.
Treat restocking as another stock movement. A cancelled order, refunded payment and received return are different events. Check the actual location quantity and release of reservations before assuming cancellation restores sellable stock.
Control timing, retries and conflicting updates
Give each stage its own completion signal so a successful initial import cannot hide later failures.
| Process | Completion signal | Recovery responsibility |
|---|---|---|
| Initial load | Every selected source record is matched to a destination record or listed as failed | Integration operator resolves mapping failures before widening the load |
| Subsequent event | The destination change is recorded against the source event and record IDs | Connector provider or custom integration team prevents duplicate processing |
| Scheduled refresh | The job records its processed boundary, accepted updates and failures | Operations checks overdue jobs and unprocessed records |
| Reconciliation | Source and destination agree for the same records and boundary | A named store operator investigates discrepancies with the system owner |
Repeated events can create duplicate external orders if the receiver treats every delivery as a new sale. Use the store identity and original Shopify order ID to recognise an order already recorded. A retry after a timeout must first establish whether the destination accepted the original request.
Out-of-order events can restore an old description after a newer one. Shopify’s webhook guidance says delivery order is not guaranteed and recommends reconciliation jobs. The integration team must compare event times or read the current record before applying an older change.
Concurrent stock changes need similar care. If Shopify sells two units while an old ERP snapshot still says 12, an uncoordinated refresh can restore stock already sold. Import and allocate the order before publishing the corresponding authoritative quantity, or use the connector’s supported conflict controls.
For custom calls, Shopify’s idempotency guidance applies to specific supported mutations. Reusing the same idempotency identifier tells Shopify that a request repeats an earlier operation. Two separate stock receipts need separate identifiers, even when both add two units. The developer must implement this for the selected API version.
| Failure | Records to inspect | Operator and next action |
|---|---|---|
| Expired or revoked access | Failed source record, account identity and authorisation error in the connector log | Account administrator renews access. Integration operator resumes from the failed boundary |
| Throttled request | Job batch, API throttle response and retry queue | Connector provider or developer backs off and retries without re-creating accepted records |
| Unmapped variant | Source SKU, Shopify variant ID and mapping error | Catalogue operator corrects the match and replays that record |
| Failed fulfilment update | NetSuite item fulfilment, original order line and destination rejection | Fulfilment operator and connector support correct the rejected mapping or line reference |
Verify the catalogue and order answer shoppers receive
Follow the changed record through to the product or order information your customer actually receives. A matching connector log is only one checkpoint in that path.
Make an authorised description change in the system that owns it. Inspect the Shopify product page with the intended variant selected. Product-level copy must describe every variant it covers. Size, colour and other variant-specific details must match the purchasable selection.
Next, change stock at the mapped source location. Confirm the expected Shopify availability, then check the selected variant’s purchase state. The example’s 12 available units mean 12 units available for sale. Twelve units physically present with two already committed mean a different availability.
Finally, fulfil only part of a test order. Inspect its Shopify line quantities and tracking, then the customer order-status view. Confirm that the customer sees the shipment that exists and the items still awaiting fulfilment.
A downstream shopping assistant needs those same distinctions. Catalogue records describe products and availability. Store policies describe delivery and refund rules. Current order records describe what has happened to a particular purchase. A product sync cannot establish that a return was received or a refund issued.
Reconcile the connection before expanding it
Compare source and destination records for the same controlled batch, using stable IDs and selected field values. Count the expected items and orders, record every failure and explain any deliberate exclusion.
| Acceptance case | Required outcome before expansion |
|---|---|
| New variant | One correct new match, with the intended creation route and no duplicate item |
| Changed SKU | The same item retains its stable IDs and future orders match the renamed SKU |
| Sale | One destination order, correct quantities and totals, with consistent sellable stock |
| Partial fulfilment | Only shipped quantities are fulfilled and remaining quantities stay open |
| Cancellation | Both order states follow the configured route, with the intended refund outcome and the correct stock movement |
| Repeated event | The existing order or adjustment is recognised and no extra transaction appears |
When a case diverges, pause the affected flow and preserve its failed record IDs. Correct the mapping or permissions, establish whether any write already succeeded and replay only the required operation. Reconcile that batch again before resuming.
Assign an owner for alerts and overdue jobs, plus an account administrator for credential renewal. Name who approves connector upgrades and who compares records after them. Keep historical backfills separate from new orders so their totals and failures remain traceable.
Expand one boundary at a time, such as another stock location or a larger item batch. Enable each additional flow only after its own acceptance cases pass.
Give your store its own shopping agent
Customers ask in their own words. Your agent finds, compares and checks out from your catalogue, right on your site.