Insights

Shopify HubSpot Integration: What Syncs, What Breaks, and How to Choose the Right Approach

By Jayant Chaudhary · November 8, 2026

A Shopify HubSpot integration is usually bought for a simple reason. The store knows who bought what, the CRM knows who opened which email and clicked which ad, and the marketing team wants both halves in one place, so it can see which campaigns actually sell and stop sending discount codes to customers who bought at full price yesterday. Connecting the two looks easy, and in a narrow sense it is. HubSpot publishes its own Shopify integration, and it installs in an afternoon. The trouble starts a few weeks later, when somebody builds a revenue report and the numbers don’t match the Shopify dashboard, or the sales team asks why none of these orders shows up in the deal pipeline.

This article is for the store owner, ecommerce manager or ops lead who already has both systems and wants them to agree. I’ll go through what needs to move between Shopify and HubSpot, object by object, and compare the three ways of doing it (a marketplace app, an iPaaS such as Zapier, Make or Celigo, and a custom build) with an honest account of where each one breaks. After that come the data-model traps specific to this pair, a real engagement of ours about tracing ads to purchases, and a testing routine to run before you trust any of it.

What the native integration does, and what it leaves to you

HubSpot’s Shopify integration (the knowledge base calls it Shopify Data Sync) is installed from the HubSpot App Marketplace, and according to HubSpot’s documentation it’s available on any HubSpot subscription. It syncs Shopify customers with HubSpot contacts in both directions, matching on email address. Products sync both ways too, matched on product name, and companies sync both ways if the store is on Shopify Plus. Orders and abandoned checkouts travel in one direction only, from Shopify into HubSpot, where they become order and cart records. After the initial sync, HubSpot says that changes come through within ten minutes. Installing it also adds the HubSpot tracking code to the store, which matters for attribution.

The documentation is just as clear about the limits. Product variants don’t sync. Shopify metafields aren’t supported. Deleting a customer in Shopify doesn’t delete the contact in HubSpot. An order placed without an email address arrives as an order with no contact attached, and the order and cart sync won’t spot or merge duplicates. Custom field mappings need at least Data Hub Starter. None of this makes it a bad integration, and for many stores it’s the right starting point. It does mean that it copies records, and leaves the decisions about what those records mean (which one is revenue, which one is the customer, which one gets credit for the sale) to you.

What actually needs to sync, object by object

Before choosing a tool, it’s worth writing down which system owns each kind of record and which direction the data should flow. For most stores, the answer looks something like this.

ObjectDirectionSystem of recordCommon trap
Customers / contactsBoth waysShopify for identity and purchase history; HubSpot for marketing consent and lifecycle stageEmail matching creates duplicates when a customer uses a second address
OrdersShopify to HubSpotShopifyOrders land as HubSpot order records, which don’t appear in deal pipelines or deal-based revenue reports
Line items and productsShopify to HubSpot (products can sync back)ShopifyVariants don’t sync natively, so size and color never arrive
Abandoned checkouts / cartsShopify to HubSpotShopifyRecovery emails fire for carts the customer already completed another way
Refunds and cancellationsShopify to HubSpotShopifyRevenue in HubSpot stays at the original order value unless something updates or reverses it
Companies (B2B)Both ways, Shopify Plus onlyUsually Shopify for B2B buyersMatching on normalized company name merges two different firms that share a name
Marketing consentBoth ways, carefullyOne system, chosen deliberatelyAn unsubscribe in one system that never reaches the other
Ad interactions and sessionsCaptured in HubSpotHubSpotAttribution only works for visitors HubSpot can tie to a contact

The rows that cause the most damage are orders and refunds. Customers are a matching problem, and matching problems can be cleaned up. A revenue figure that has been wrong for a quarter, and on which budget decisions were made, is much harder to undo.

Native app, iPaaS or custom build

There are three ways to put a Shopify HubSpot integration together, and most stores that have been running for a few years end up with some combination of them.

ApproachWhen it fitsWhere it breaks
HubSpot’s own integration, or another marketplace appStandard objects, one store, marketing wants contacts, orders and carts in HubSpot quicklyOrders stay out of deal pipelines and deal-based attribution; no variants or metafields; limited control over matching and transformation
iPaaS (Zapier, Make, Celigo and similar)Specific gaps on top of the native sync, such as creating a deal from a paid order or pushing a tag back to ShopifyLogic spreads across dozens of small flows nobody owns; retries and duplicates are easy to get wrong; per-task pricing can climb with order volume
Custom build on the Shopify and HubSpot APIsSeveral stores, B2B and retail on one portal, attribution that has to hold up in a board meeting, or data from an ERP joining inCosts more up front and needs an owner, monitoring and someone who understands both APIs

The native integration deserves to be tried first, because it covers a surprising amount and is cheap to evaluate. HubSpot’s App Marketplace and the Shopify App Store also list third-party connectors that fill some of its gaps. The judgment you have to make is whether your requirements are a short list of gaps or a different data model altogether. A store that needs an extra property on the contact is in the first group. A store that wants every order to count as attributed revenue, net of refunds, across two storefronts, is in the second, and stacking automations on top of the native sync won’t get it there reliably.

On the custom side, one point is worth knowing before anyone writes code. Shopify has treated its REST Admin API as legacy since October 1, 2024, and new public apps have had to use the GraphQL Admin API since April 1, 2025. Anything built today should be built on GraphQL, which is also where Shopify’s cost-based rate limits apply, with bucket sizes that vary by plan.

The traps specific to Shopify and HubSpot

Orders and deals are different objects

This one catches almost everybody. In HubSpot’s data model, Shopify orders become records of the orders object, which HubSpot describes as the operational record of what was purchased and how it will be fulfilled. Deals, invoices and payments are where HubSpot keeps pricing and revenue, and an order can be associated with a deal, but the integration doesn’t create one. So the pipeline board stays empty, and any report built on deals ignores the store entirely.

The usual fix is a dedicated ecommerce pipeline, kept well away from the sales pipeline, with a deal created for each order once it reaches an agreed status (paid, say, or fulfilled). That works if someone decides the rules in advance. Which status creates the deal? What happens to the deal when the order is refunded, partially refunded or cancelled? Is the deal amount the gross total, the total net of discounts, or the total net of tax and shipping? These are accounting questions dressed up as CRM settings, and if nobody answers them, the integration answers them for you, usually inconsistently.

Attribution runs on deals, and on identified contacts

HubSpot’s revenue attribution reports require Marketing Hub Enterprise, and its documentation describes them in terms of closed-won deals with at least one associated contact, using the deal’s amount. Orders don’t feature. So a store that wants HubSpot to say which campaigns and ads produced revenue needs deals that represent orders, with the right contact associated, before those reports can say anything meaningful.

The other half of attribution is identity. HubSpot can only connect an ad click or a website session to a purchase if it knows that the visitor and the buyer are the same person, and that usually depends on the tracking code, a cookie and an email address that matches. A shopper who browses on a phone and buys on a laptop, or checks out with a different email, breaks the chain. No integration fixes that completely.

One person, several records

The native sync matches contacts on email, which is a sensible choice that still produces duplicates in practice. Customers use a work address for one order and a personal one for the next. HubSpot’s documentation for the integration also notes that orders from customers without an email address arrive without a contact. Decide early which identifier wins and who reviews merges, because every duplicate splits a customer’s purchase history and weakens the segmentation you connected the systems for.

Refunds, currencies and the things the documentation doesn’t mention

HubSpot’s documentation doesn’t say how refunds, partial refunds or currency are represented on synced orders, so check it in your own portal with real test orders before any report depends on it. On the Shopify side, money fields in the GraphQL Admin API carry both a shop-currency amount and a presentment-currency amount, so a store that sells in several currencies has to decide which one HubSpot reports in, and say so in writing.

Customer data access

If you build a custom app, Shopify treats customer names, emails, phone numbers and addresses as protected customer data, and an app has to be approved for each of those fields. Fields that aren’t approved come back redacted, as nulls with an error, even though the request itself succeeds. Better to discover that on day one than in production, when contacts arrive in HubSpot with no names.

Tracing ads to purchases: one engagement

Through an agency partner, we connected Shopify and HubSpot for a fast-growing Atlanta retailer of sports apparel and athletic shoes. The team wanted three things from the same data: to trace their ads to the purchases those ads produced, to retarget the right customers, and to serve buyers better once they had bought.

Those three goals explain why a project like this is more than a connector. Tracing an ad to a purchase means the purchase has to land in HubSpot attached to the right contact, in a form HubSpot’s attribution can read. Retargeting the right customers means the audiences sent to the ad networks have to be built on clean purchase data. And serving buyers better means that whoever answers a customer’s question can see what that customer actually ordered. All three rest on the same foundation, a single trustworthy picture of who bought what.

How to test and reconcile before you trust it

A Shopify HubSpot integration should earn trust with evidence, which is cheap to produce. Before go-live, place a set of test orders that covers the cases you actually have: a new customer, a returning customer with a second email address, a guest checkout, a discount code, a multi-currency order if you sell abroad, a partial refund, a full refund and a cancellation. Then check each one in HubSpot by hand – the contact it attached to, the amount, the products, and the deal, if you create deals – and write down what you expected next to what you got.

Once it’s live, reconcile on a schedule. Count orders and sum revenue in Shopify for a fixed window, do the same in HubSpot, and investigate any gap that exceeds a small agreed tolerance. Shopify’s developer documentation makes the same point about webhooks: delivery isn’t guaranteed, duplicates can arrive and order within a topic isn’t guaranteed, so it recommends periodic reconciliation jobs that fetch recently updated records. If you’re building custom, make the handlers idempotent from the start, keyed on the Shopify webhook or order ID, so that a retry never creates a second deal. And keep a duplicate-contact report running, because duplicates are far cheaper to merge weekly than annually.

Related reading

How we approach it at eProxim

eProxim has been engineering integrations since 1997, from Alpharetta, Georgia, with delivery centers in Iași, Romania and Gurgaon, India, and we’ve built ecommerce stores for more than 25 years. Shopify HubSpot work usually starts with the native integration and a list of things it doesn’t do, so we begin with the questions underneath: what should count as revenue, who the customer is, and which system owns which record. Sometimes the answer is a few additions to the native sync, and sometimes a custom layer that treats orders, refunds and attribution as one model. Either way, it should be the smallest thing that gives you numbers you can defend.

We can usually start within a week, small fixed-scope jobs are welcome, and we work by fixed price, by the hour or as a monthly block of hours. If you’re an agency with a client who needs this, our Shopify partner page explains how we work behind you, and our page for small agencies and dev shops sets out the terms. If you want Shopify and HubSpot to agree, tell us what you’re working with.

About the author

Jayant Chaudhary

Jayant Chaudhary is a technology executive, entrepreneur, and recovering optimist about how businesses make decisions. After more than 30 years in the industry, he's learned that technology is rarely the hardest part. People, politics, and PowerPoint usually are. He writes about business, technology, leadership, and lessons learned the expensive way. He has strong opinions, but reserves the right to change them when confronted with facts. This, as we all know, is an increasingly unfashionable habit.

Working with eProxim

Have integration work your team can’t - or doesn’t want to - take on? We deliver it under your brand.