Shopify customer retention, and the cohort report's blind spot

What Shopify customer retention actually measures

Shopify customer retention is the share of customers who buy from your store again after a first order, measured over a set window and reported in Shopify by the cohort in which they first purchased. Retention gets treated as something you add after the sale: points, tiers, email flows, a subscription. The most ordinary reason a customer doesn't come back is duller than any of that. Their last order broke, and nobody fixed it before they noticed. An order is marked fulfilled, the label prints, and the tracking number never scans. The customer waits nine days, opens a ticket, and gets a refund on day eleven. That refund closed the ticket. It didn't buy the second order, and the cohort report counts the customer as churned without saying why.

Retention and satisfaction are also different numbers. Satisfaction asks whether a customer said they were happy. Retention asks whether they came back, and only one of the two can be banked.

Where the number lives in Shopify, and what it will not tell you

Shopify computes this for you already, inside Shopify's reports and analytics. The customer cohort analysis report groups customers by the date of their first order, and its metric menu switches the view to customer retention rate. The blunter split sits in the new versus returning customers report, where a first-time customer is one who has placed a first order and a returning customer is one whose history already includes at least one.

1. Open Analytics, then Reports, in your Shopify admin.
2. Open the customer cohort analysis report.
3. Set the metric menu to customer retention rate.

Now read what it doesn't say. The cohort report tells you which month's customers stopped coming back. It carries nothing about whether those orders were fulfilled correctly, so the one segmentation that would explain a bad cohort is the one the report doesn't offer. You can split retention by acquisition channel, by first product, by location. You can't split it by whether the promise made at checkout was kept.

The five metrics everyone lists, and the one nobody segments

Open any retention guide and you meet the same five numbers. They're the right five, and every one of them is an outcome. I add a sixth.

• Repeat customer rate: the share of customers who have placed more than one order.
• Purchase frequency: how many orders the average customer places in the window.
• Average order value: revenue divided by orders, and the number most easily flattered by a promotion.
• Customer lifetime value: what a customer is worth across the whole relationship, not a single basket.
• Churn rate: the share who didn't come back inside that window.
• Retention by order outcome: the same retention rate, split by whether that customer's first order shipped clean or hit an operational break. It's the only one of the six that names a cause you can fix rather than a result you can report.

The first five describe what happened. The sixth describes why. Reichheld and Sasser made the original case for retention as a profit lever in Zero Defections, and the argument was about defects, not discounts: find what drives customers away and remove it. A retention program built on the first five alone optimizes for the customers you already kept, which is why generic customer retention strategies stall on a store whose real problem is fulfillment.

Why a broken order costs you the next order

Every order is a promise: this item, at this price, by this date. A break is that promise failing in the window where the customer is paying you the most attention they ever will.

What customers punish isn't usually the delay itself. A study of 466 cross-border shoppers found that delivery information service, the updates telling a buyer where their order is, significantly drove satisfaction while delivery service quality on its own did not, and satisfaction drove repurchase intention. Read that as an operator: they forgive the late parcel and remember finding out late. Baymard's testing agrees from the interface side. Order tracking is the most important account feature for 50% of respondents, yet 67% of tested sites don't consistently provide the key tracking details and 25% fail to show a reliable delivery date.

There's a legal floor under this too. The FTC's Mail Order Rule requires you to ship inside the time you advertised, or within 30 days where you advertised none, and on a delay to seek consent, offer cancellation and refund promptly. A late order is an obligation before it's a retention problem, and your on-time delivery rate won't tell you which orders are about to breach it.

What actually breaks after checkout on a Shopify store

One in five orders hits an operational break after checkout. Shopify records almost all of them accurately: the fulfillment order status can read OPEN, IN_PROGRESS, ON_HOLD, INCOMPLETE, SCHEDULED, CLOSED or CANCELLED, so the state of an order is never a mystery. Recording isn't resolving.

• Stuck fulfillment order: status sits at OPEN past your ship window. The customer got a confirmation email and then silence.
• On hold: oN_HOLD after a payment or inventory check nobody cleared, invisible to a buyer who believes the order is moving.
• Partial or split shipment: iNCOMPLETE, or two parcels and one explanation, so half the order looks lost.
• Fulfilled but never scanned: marked fulfilled with a tracking number that never moves, which reads to the customer as a lie.
• Address or delivery exception: the carrier stops, nobody upstream reacts, and the first person to notice is the buyer.
• Return that stalls: approved, then silent, exactly when the customer is deciding whether to shop with you again.

That last one isn't an edge case. NRF puts 19.3% of online sales on the returns leg in 2025, and Baymard, benchmarking the returns flows of the 42 largest e-commerce sites, found 54% with substantial returns usability issues. Every break above is visible in your data before the customer feels it, which is why Shopify order tracking starts too late.

Detect, decide, act: retention as an operations job

The signal already exists. Shopify emits webhooks for order and fulfillment order events as they happen, so a hold or an order that stopped moving is observable in real time by anything listening. Something still has to listen, decide the right action, and take it. That gap is the whole argument. We detect the break from the signal the platform already emits, decide what should happen by measuring the order against the promise made at checkout, then act, usually before the customer knows anything went wrong.

A helpdesk is a system for replying about problems, not resolving them, and answering faster about a stuck order doesn't unstick the order. That's not a knock on the tool, it's the wrong category for the job. The right one is proactive e-commerce operations, a different function rather than a better inbox.

The edges matter. Keeyu isn't a loyalty app, not an email or SMS platform, not a subscription tool, not a reviews platform, not a returns portal, not a helpdesk, not a carrier and not an OMS. We don't run your returns process, we catch the return that stopped moving inside it. Your loyalty program and your email flows keep their job. They work on the customers you kept. We work on the reason customers leave.

Where to start, in the order that pays

This is a sequence, not a strategy list.

1. Pull the customer cohort analysis report and find your worst cohort.
2. Separate that cohort's first orders into the ones that shipped clean and the ones that did not.
3. Compare the repeat rate of the two groups.
4. Fix the detection before you fund the campaign.

The difference between those groups is your retention leak, and unlike a benchmark it has a cause you can act on. If the gap is small, your fulfillment is healthy and the tactics everyone else writes about are the right next move. If it's large, no loyalty program will close it, because the customers you're paying to win back are the ones you already failed once.

Every order is a promise

If your cohort report keeps telling you which month went bad and never why, the answer isn't another campaign. It's the orders that broke after checkout and were never fixed. Keeyu watches every Shopify order against what it was promised, catches the ones that break, and acts on them, usually before the customer feels it. Every order is a promise, and retention is what happens when you keep it. See a demo.

• Shopify Help Center. Customers reports. The customer cohort analysis report, its metric menu, and the first-time versus returning customer definitions.
• Shopify Help Center. Shopify reports and analytics. Where retention reporting sits in the admin.
• Shopify. FulfillmentOrderStatus, GraphQL Admin API. The valid statuses a fulfillment order can hold: OPEN, IN_PROGRESS, ON_HOLD, INCOMPLETE, SCHEDULED, CLOSED and CANCELLED.
• Shopify. Webhooks. Order and fulfillment order events are emitted as they happen, so a break is observable in real time.
• Baymard Institute. Always Provide 6 Key Order-Tracking Details. Order tracking is the most important account feature for 50% of respondents, 67% of test sites do not consistently provide all key tracking details, and 25% fail to reliably provide an expected delivery date.
• Baymard Institute. Order Returns Experience and Customer Retention. 54% of sites have significant usability issues in the returns flow.
• National Retail Federation and Happy Returns. 2025 Retail Returns Landscape. An estimated 19.3% of online sales returned in 2025, against total returns of $849.9 billion.
• Federal Trade Commission. Mail, Internet, or Telephone Order Merchandise Rule. The shipping window, the 30 day default, and the seller's duty to seek consent, offer cancellation and refund promptly on a delay.
• Reichheld, F. F. and Sasser, W. E. Zero Defections: Quality Comes to Services. Harvard Business Review, September 1990. The original case for treating customer defection as a quality defect with a profit consequence.
• Hui, T., Al Mamun, A., Reza, M. N. H. and Wan Hussain, W. M. H. Logistic service quality and cross-border repurchase intention. Heliyon, 2025. 466 respondents: delivery information service significantly drives satisfaction (beta 0.231, p below .001) where delivery service quality alone does not (beta 0.057, not significant), and satisfaction drives repurchase intention (beta 0.305, p below .001).