Shopify integrations with ERP, accounting, CRM and marketplaces

A Shopify integration moves data between your store and the other systems your business runs on: accounting, ERP, CRM, helpdesk, marketplaces, payment providers and AI services. For simple set-ups an App Store connector is enough. When your mapping, volume or edge cases are specific to your business, you need a custom integration app built on the GraphQL Admin API and webhooks, with proper retry and reconciliation.

Each guide below covers what data flows between the two systems, the off-the-shelf options, where they break, and how a custom integration is designed.

• NetSuite integrationHow a Shopify NetSuite integration works: order, customer, item, inventory, fulfillment, refund and payout flows, where connectors like Celigo break, and how a custom integration is designed.
• QuickBooks integrationHow to integrate Shopify with QuickBooks Online or QuickBooks Desktop: per-order vs summary accounting, payouts, fees, refunds and sales tax, where connectors fall short, and what a custom integration does.
• Xero integrationHow to integrate Shopify with Xero, written for UK and Australian merchants: VAT and GST on tax-inclusive prices, payouts in GBP and AUD, fees, refunds and multi-currency, and when to build a custom Xero integration.
• HubSpot integrationHow a Shopify HubSpot integration works: customers, orders, products and carts in HubSpot, lifecycle stages, B2B companies and deals, where the standard app stops, and how to build a custom sync.
• Salesforce integrationHow to integrate Shopify with Salesforce Sales Cloud and Service Cloud: accounts, contacts, orders, products and cases; connector and middleware options; and a custom integration design with upserts and external IDs.
• Amazon integrationHow to integrate Shopify with Amazon: listing sync, order import, shared inventory, FBA and multi-channel fulfillment, Buy with Prime, where connectors break, and when a custom SP-API integration makes sense.
• ERP integrationHow Shopify ERP integrations work with SAP, Microsoft Dynamics 365, Acumatica and Odoo: which system owns what, order, inventory, pricing and fulfillment flows, connector vs middleware vs custom app, and a decision checklist.
• ChatGPT & AI integrationHow to build a Shopify ChatGPT or OpenAI integration properly: AI product copy, support answers from real order data and smarter search, with data privacy, human review and cost control designed in.
• Stripe & Square integrationWhat a Shopify Stripe integration or Square integration can and can't do: checkout payment options, moving Stripe subscriptions to Shopify, syncing Square POS with Shopify, and reconciling payouts across providers.
• eBay, Etsy & Walmart integrationHow to sell on eBay, Etsy and Walmart from Shopify: listing sync, order import, shared inventory and marketplace-specific rules, which connectors exist, where they break, and when to build a custom marketplace integration.
• Zendesk integrationHow a Shopify Zendesk integration works: order, fulfillment and customer data inside tickets, refunds and cancellations from Zendesk, B2B context, where the standard app stops, and how to build a custom Zendesk sidebar app.

The three ways to integrate Shopify

1. A connector app. Many systems publish their own Shopify app, and third-party connector apps cover most popular pairs. You install it, map a few fields and it runs. This is the right starting point for most stores. Connectors are built for the common case, though. They work well until your process isn't the common case.

2. An iPaaS or middleware platform. Integration platforms such as Celigo, Workato, MuleSoft or Boomi, and lighter tools like Zapier or Make, give you pre-built Shopify connectors plus a place to write your own mapping and flows. They are good when you connect several systems and have someone who can own the flows. You still have to design the edge cases. The platform only gives you somewhere to put them.

3. A custom integration app. A Shopify app built for your store that subscribes to webhooks, calls the GraphQL Admin API and talks to the other system's API directly. It costs more upfront, but it does exactly what your process needs, and the logic lives in code you own and can test.

These options mix. A common, sensible set-up uses the accounting system's own connector for payouts, an iPaaS for CRM sync, and a small custom app for the one ERP flow that nothing else handles.

Where connectors usually break

The same problems come up across almost every system:

• Refunds and order edits: partial refunds, refunds without restock, exchanges, and edits after an order has already been exported.
• Multi-location inventory: stock per warehouse, transfers, and which location a fulfillment ships from.
• Tax: different tax codes per country or state, tax-inclusive pricing for UK, EU and AU stores, and duties.
• Currencies: orders in the customer's currency, payouts in the store's currency, and the FX difference in between.
• Bundles and kits: one Shopify line item that is three SKUs in the ERP, or the reverse.
• B2B: company accounts, payment terms, customer-specific prices and PO numbers.
• Volume: a sync that works at 50 orders a day falls behind at 5,000 during a sale.
• Silent failures: a record fails to sync and nobody knows until month-end.

How a custom Shopify integration is designed

Every integration guide on this site describes the same core architecture, adapted to the other system:

1. Events in. The app subscribes to Shopify webhooks such as orders/create, orders/updated, refunds/create, products/update and inventory_levels/update. Each delivery is verified with its HMAC and written to a queue immediately, and the app returns 200 right away.
2. Idempotent processing. A worker takes each event and writes to the other system, using the Shopify ID as an idempotency key or external reference. Shopify can deliver a webhook more than once, so processing the same event twice must not create a second invoice or sales order.
3. Retry with back-off. If the other system is down or rate-limited, the event is retried later. After a set number of attempts it goes to a dead-letter list with the error message, visible to staff.
4. Pull where push isn't possible. For data that changes in the other system, such as inventory in the ERP or payment status in accounting, the app either receives that system's webhooks or polls on a schedule and writes to Shopify through the GraphQL Admin API. Large updates use bulk operations.
5. Reconciliation. A scheduled job compares both sides for a period: order counts, totals and stock levels. It flags or repairs differences. This is what makes an integration trustworthy.
6. Traceability. Every record stores the IDs from both systems, in Shopify metafields and in the other system's reference fields, so anyone can trace an order from one side to the other.

Decision checklist

• Which system is the source of truth for products, prices, inventory and customers?
• Which direction does each data type flow, and how quickly must it arrive?
• What are your peak order volumes and SKU count?
• Which edge cases happen in your business: refunds, exchanges, bundles, B2B, multi-currency?
• Does the other system have a documented API and a sandbox?
• Who will look at failed syncs, and where?
• What does a wrong or missing record cost you: a support ticket or a tax filing?