Connect Mailchimp to Shopify: campaign data into the store
The return direction: attributing orders to campaigns, writing subscriber status back to customer records, and triggering store actions from email behaviour.
Rebecca Pearson10 min read
On this page
Most integration guides cover store data flowing into Mailchimp, which is the obvious direction and the one the official connector handles. The return flow gets much less attention and is where a good deal of the commercial value sits: knowing which campaigns produced which orders, keeping subscriber status accurate in your customer records, and letting email behaviour trigger things in the store.
This page covers that direction. For getting customers, orders, and segments into Mailchimp in the first place, see connecting Shopify to Mailchimp.
What we'll cover
Why the return direction matters
Three concrete problems are solved only by data moving back into the store.
Attribution lives in the wrong place. Mailchimp reports revenue per campaign in Mailchimp. Your store reports orders in Shopify. When someone asks which channel drove the quarter, those two numbers sit in different tools with different definitions, and reconciling them by hand is the monthly task nobody enjoys.
Subscriber status goes stale in the store. Somebody unsubscribes in an email, and Shopify continues believing they accept marketing. The next sync re-adds them, they receive another campaign, and you have a complaint and a compliance problem from a synchronization gap.
Email behaviour never reaches the people who could use it. A customer opening every campaign and clicking through repeatedly without buying is a useful signal for whoever handles sales or support, and it stays locked in the email tool where nobody looks.
Attributing orders to campaigns
Getting attribution onto the order record itself is what makes it usable everywhere else.
Capture the source at checkout. Campaign links carrying UTM parameters can be preserved through the session and written to the order as a note or metafield. This is the foundation, and getting it right at the point of sale is considerably easier than reconstructing it afterwards.
Reconcile against Mailchimp's own reporting. Mailchimp attributes revenue by its own rules, and your UTM capture will disagree with it. Both numbers are defensible and they measure slightly different things. Comparing them and understanding the gap is more useful than picking one and hoping nobody checks.
Decide your attribution model deliberately. Last click is simplest and undervalues campaigns that started the interest. First touch does the opposite. Pick one, write it down, and apply it consistently, because the argument about which is correct is much less costly than the confusion of different people assuming different models.
Handle the multi-touch case. Someone receiving three campaigns before buying is normal. Storing all touches on the order rather than only the last preserves the option of changing the model later without losing history.
Write it where reporting can reach it. An order metafield or tag makes campaign attribution available to Shopify reports, exports, and anything else reading orders. Attribution that exists only in a separate dashboard gets used far less.
Writing subscriber status back
The highest-value part of this direction, and the one with compliance consequences if omitted.
Unsubscribes must flow back promptly. When someone unsubscribes in Mailchimp, the corresponding Shopify customer's marketing consent should be updated. Without this, your next sync treats them as subscribable and the cycle repeats. This alone justifies building the return path.
Handle every status, not just unsubscribed. Cleaned addresses, bounces, and complaints each mean something different. A hard bounce means the address is dead and should stop being used anywhere, including for order notifications, which is worth knowing at fulfilment time. A spam complaint is a stronger signal than an unsubscribe and deserves recording as such.
Record when and where. The status and the timestamp, so you can evidence what happened. Where regulation requires demonstrating consent, the audit trail matters as much as the current state.
Keep transactional email separate. Someone unsubscribing from marketing should still receive their shipping notification. If your status write-back suppresses all email to that customer, you have broken order communications, which is a worse problem than the one you solved.
Make it near real time. The window between unsubscribing and the store knowing is the window in which you can send them another campaign. A nightly sync leaves a day of exposure; an event-driven update closes it.
Triggering store actions from email behaviour
Once email events reach the store, several useful patterns become available.
Tag engaged customers. Consistent openers and clickers who have not bought recently are a receptive audience for a targeted approach, and tagging them in Shopify puts that in front of whoever works the customer list.
Flag disengagement before it becomes churn. A customer who read every campaign for a year and has opened nothing for three months is signalling something. Surfacing that in the store, where account management happens, gives someone the chance to act.
Generate a discount on a specific action. Clicking a particular link can trigger creation of a single-use discount code, which is more controlled than sending the same code to everybody and lets you attribute the redemption precisely.
Alert on high-value engagement. A customer above a spend threshold clicking a product link is worth a notification to whoever handles that relationship, particularly in businesses with a sales motion alongside the store.
Suppress campaigns during an open issue. A customer with an unresolved complaint should probably not receive a promotional email that week. This needs the support system in the loop as well, and it is the kind of judgement that separates careful operations from automated ones.
Each of these is straightforward to describe and fiddly to assemble from webhooks and API calls. On CodeWords you describe the behaviour you want in plain language: which email events matter, what should happen in the store, and what conditions should prevent it. Cody, the automation builder, builds it, connects it to Mailchimp and Shopify, and deploys it. Automations connect to more than 3,000 integrations, which covers the support and analytics tools these rules often need to consult. The free plan covers light use, with Pro at $39 per month and Business at $100 per month as usage grows; details are on the pricing page.
Bringing email data into store reporting
Beyond individual customer records, aggregate email data is worth having alongside store data.
Campaign performance next to actual orders. Sends, opens, clicks, and the orders that followed, in one place, using your own attribution rather than the email tool's. This is the report that actually answers whether email is working.
List health over time. Audience size, growth rate, unsubscribe rate, and bounce rate tracked as a series. A rising unsubscribe rate is an early warning about sending frequency or relevance, and it is visible long before revenue moves.
Revenue per subscriber. Total email-attributed revenue divided by audience size, tracked over time. A useful check on the instinct to grow the list at any cost, since a larger list with lower engagement can produce less.
Segment performance. Which segments respond and which do not, which tells you whether your segmentation is doing anything. Segments that perform identically to the whole list are not earning their complexity.
What to be careful about
Sync loops. A change in Shopify updating Mailchimp, which triggers an update back to Shopify, which triggers Mailchimp again. Mark automated changes so each side can recognize its own updates and ignore them.
Rate limits on both sides. Bulk operations against either API will hit limits. Batch, back off, and resume rather than retrying immediately.
Identity mismatches. Email is the join key and it is not reliable: people change addresses, use different ones at checkout, and share them. Decide a rule for reconciling and accept it will be imperfect.
Timing gaps. Attribution requires the order and the campaign data to arrive close enough together to associate. Late-arriving data needs a plan, usually a reconciliation pass rather than a real-time join.
Personal data moving between systems. Every field you copy is a field in another place, subject to deletion requests and retention policy. Copy what the process needs rather than everything available, which is easier to justify and easier to delete.
A sensible build order
The return direction has several independent pieces, and building them in order of value gets the important one working first.
Unsubscribe write-back, first and alone. It carries compliance consequences, it is the smallest piece, and it works without anything else being in place. Get it running near real time and confirm it with a test unsubscribe before moving on.
Bounce and complaint handling next. The same mechanism extended to the other status types. A hard bounce flagged on the customer record helps fulfilment as well as marketing, which makes it easy to justify.
UTM capture at checkout. Start collecting attribution before you need to report on it, since this is the one piece you cannot backfill. Even if nothing reads the field for a quarter, the data accumulates from the day you switch it on.
Attribution reporting once there is history. With a few months of captured sources, the campaign-to-order reporting becomes meaningful. Building the report before the data exists produces an empty dashboard and a loss of interest.
Behavioural triggers last. Engagement tagging, discount generation, and alerting are the most interesting and the least essential. They also depend on judgement about what should suppress them, which is easier to specify once the simpler flows have shown you how the data actually behaves.
Frequently asked questions
Does the official connector handle the return direction?
Partly. It reports campaign revenue within Mailchimp and handles some status synchronization. Attribution written onto the order record, email behaviour triggering store actions, and combined reporting generally need building.
How quickly should unsubscribes reach Shopify?
As close to immediately as you can manage, since the delay is your exposure window for sending another campaign to someone who has opted out. Event-driven updating rather than a scheduled sync is the right shape here.
Why does my attribution disagree with Mailchimp's?
Because the two use different rules about which campaign gets credit and over what window. Neither is wrong. Understand the difference, pick the one you will report on, and keep the other as a cross-check.
Can I trigger a Shopify discount from an email click?
Yes. Click events can create a single-use code, which is more controlled than broadcasting one and makes redemption attributable to the individual rather than to the campaign in aggregate.
What should happen when an address hard bounces?
Treat it as invalid everywhere rather than only in marketing. An address that bounces will also fail for order confirmations, so flagging it in the store lets someone obtain a working address before a delivery problem occurs.
Is this worth building for a small store?
The unsubscribe write-back is worth it at any size, because the compliance risk does not scale with revenue. The attribution and behavioural triggers pay off once email is a meaningful channel, which is usually when campaigns start driving a noticeable share of orders.
How do I test this without emailing real customers?
Use a test audience with addresses you control, and trigger each status change by hand: unsubscribe, bounce, complaint. Confirming each one arrives correctly in Shopify takes twenty minutes and is considerably cheaper than discovering a gap through a customer complaint.
Does this work with other email platforms?
The patterns transfer directly, since every email platform exposes subscriber status and engagement events in some form. The field names and event structures differ, and the decisions about attribution, consent, and suppression are identical.
Should campaign attribution go on the order or the customer?
The order, since a customer places many orders from different sources and collapsing that to one field on the customer loses the detail that makes attribution useful. A summary on the customer record is a reasonable addition on top, not a replacement.