A skincare brand launches a subscribe-and-save program on its best-selling serum, and three months later the founder pulls up Shopify's Analytics > Reports and sees the returning customer rate has climbed from 31% to 74%. It reads like a turnaround - the kind of number that gets screenshotted into a pitch deck as evidence the product finally has real loyalty behind it. Nobody on the team is misreading the dashboard. The report is doing exactly what it's built to do: counting every customer who placed more than one order in the period as returning. What it can't do is tell the difference between a customer who thought about the brand and chose to buy again, and a subscriber whose card gets charged automatically every month whether she thinks about the brand at all.
Before the subscription program existed, a returning customer on that dashboard meant something specific: someone who ran out of serum, remembered the brand, and went back to the site to buy it again - a real, repeated decision. Once auto-renewal is running, most of the "returning" orders behind that 74% never involved a decision at all. The subscriber signed up once, and Shopify's Order object fires again every cycle on schedule, indistinguishable in the report from a customer who woke up one morning and chose to reorder. The report hasn't changed. What counts as "returning" underneath it has.
The mistake isn't running the subscribe-and-save program, and it isn't checking Shopify's built-in returning customer rate - both are fine things for a store to do. The mistake is reading a number that was designed for a world of one-time purchases as if it still means the same thing after most of the repeat orders behind it are mechanical instead of chosen.
Why a renewal and a reorder look identical to Shopify's own reports
- Shopify's returning customer rate and its new-vs-returning sales split both key off order count per customer in the period - they have no field that distinguishes an order created by a selling-plan renewal from one created by a customer manually checking out again
- A subscriber who renews for eighteen straight months contributes eighteen "returning customer" orders to the report without a single one of them being a choice made in that moment - the choice was made once, at signup, and the report has no way to mark the other seventeen differently
- Sales by customer type attributes 100% of every renewal's revenue to "returning customer sales," which inflates the repeat-revenue side of that split independent of whether the subscription base is healthy, growing, or one bad cycle from churning
- The distortion only grows in one direction - it never corrects itself, because a subscriber who never cancels keeps generating fresh "returning" orders for as long as the card keeps working, long after the report should have any doubt about calling it loyalty
- None of this is a bug in Shopify's reporting - the report was built and named before recurring billing was common on the platform, and it still answers the question it was built to answer: did this customer place more than one order. It just isn't the question a store with a subscription base actually needs answered anymore
A returning customer rate built for one-time purchases doesn't break when you add subscriptions. It just starts answering a question nobody's asking anymore, while looking exactly like it always did.
What trusting the blended number actually costs
The skincare brand's 74% doesn't cause a visible failure - it causes a series of reasonable-looking decisions built on a number that's mostly measuring auto-renewal instead of affection. A retention team that sees the rate climbing eases off win-back campaigns and lifecycle emails for one-time buyers, reasoning that repeat behavior is already strong, when the one-time-buyer cohort underneath the blended number may not have moved at all. A board update leans on the rate as evidence of product-market fit for a story about loyalty, when the honest version of that story is closer to "subscribers keep getting billed unless they take action to stop it." And a marketing team benchmarking against industry repeat-purchase figures - numbers built almost entirely from one-time-purchase brands - ends up comparing an inflated, subscription-heavy rate against a benchmark measuring something else entirely, drawing conclusions from a comparison that was never apples to apples.
The report isn't lying. It's answering the same question it always has - did this customer order more than once - on a business where most of the second orders now happen without anyone deciding to place them.
How to calculate a repeat-purchase rate that still means something
- Split the customer base into two cohorts before running any repeat-purchase math: customers whose orders in the period are entirely one-time purchases, and customers with at least one subscription renewal in the period
- Calculate Shopify's standard returning customer rate only against the one-time-purchase cohort - that's the number that still means "chose to come back," the same thing it meant before the subscription program existed
- For the subscription cohort, track a different, subscription-specific number instead: active-subscriber retention by cohort month, which answers whether subscribers are staying past their first few cycles, not whether their renewal orders count as repeat purchases
- When a subscriber's subscription lapses or gets canceled and she later places a genuine new one-time order or restarts a plan on her own, count that as a real returning-customer event - the distinction is whether a decision happened in that order, not which product she bought
- Report the two numbers side by side to the board or the team, not blended into one - a one-time-purchase repeat rate of 28% next to a 12-month subscriber retention rate of 61% tells a more honest story than a single 74% that quietly averages a mechanical process with a chosen one
Where this lives in AppFox Subscription
Subscription analytics, on the Growth plan and above, reports renewal, skip, pause, and cancellation events per subscriber separately from one-time order activity, which is exactly the split Shopify's own returning-customer report doesn't make - it's the raw material for running the one-time-purchase cohort and the subscriber cohort as two different calculations instead of one blended rate. Active-subscriber counts by signup cohort are available from the same dashboard, so a subscriber-retention curve doesn't have to be reconstructed from a raw order export before anyone can compare it to anything.
What AppFox doesn't do is relabel or override Shopify's own Analytics > Reports - the built-in returning customer rate and new-vs-returning sales split still blend subscription and one-time orders together exactly as Shopify built them, because that's Shopify's report, not AppFox's. What AppFox's subscription analytics gives a merchant is the subscriber-specific data needed to build the second number next to it - the one that separates a chosen reorder from a renewal that simply hadn't been stopped - so a dashboard number gets read for what it actually measures instead of what it happens to resemble.
The skincare brand didn't need to distrust its Shopify dashboard or stop checking the returning customer rate - it needed a second number next to the 74%, one that only counted the orders where a subscriber actually made a choice. That number came in closer to 40% once subscription renewals were set aside, still a real improvement worth reporting, just not the same story the blended figure had been quietly telling for three months.