If you build or maintain ecommerce systems, you have probably watched a team pour engineering effort into checkout conversion while the post-purchase side runs on a single order confirmation email. This is the case for why the code you write after checkout is worth more than the code you write before it.
The Unit Economics Nobody Puts on the Dashboard
Run the arithmetic on an ordinary store. Acquisition costs $40. The customer places a $65 order at 40 percent gross margin, so you clear $26 against $40 spent. That customer is $14 underwater the moment the box ships.
The first order is not the transaction, it is the purchase of a relationship. The second order is where margin appears, because the same $65 delivers the full $26 with nothing subtracted for acquisition.
This is why Bain's finding that a 5 percent increase in retention moves profit 25 to 95 percent is not marketing hyperbole. It is arithmetic on a base that starts negative. Returning customers spend roughly 67 percent more per order, and by the fourth purchase average order value is about 20 percent above the first.
Meanwhile acquisition inflates. Paid costs have climbed 15 to 30 percent a year since 2022, and ecommerce cost per acquisition on Meta now exceeds $45. Selling to an existing customer succeeds 60 to 70 percent of the time versus 5 to 20 percent for a cold prospect.
Five Numbers That Actually Track Retention
Most analytics dashboards report revenue and sessions, which tell you nothing about whether the base is compounding. The five that do:
Customer retention rate. ((customers at end - new customers) / customers at start) x 100. Typical ecommerce lands between 25 and 40 percent annually. Above 50 percent is top decile.
Repeat purchase rate. Customers with two or more orders divided by unique customers. Benchmarks run 25 to 30 percent for most verticals, 40 to 50 percent for consumables, and 15 to 20 percent for big-ticket items.
Customer lifetime value. Average order value times purchase frequency times average lifespan. At $55, 3.2 orders a year, and 2.5 years, CLV is $440. That number is your acquisition ceiling.
Purchase frequency. Total orders over unique customers for a period. Rising frequency against stable retention is the healthiest signal you can get.
Cohort churn. Churn segmented by acquisition month. Aggregate churn hides the thing you need to know, which is whether recent cohorts retain better than old ones. If they do not, the product or the experience is not improving, no matter what the roadmap says.
Where Automation Beats Another Ad Dollar
Every one of these is a scheduled job and a query away, which is what makes retention an engineering problem more than a budget problem.
Post-purchase sequences fire on order state: confirmation, shipping with tracking, a delivery follow up, then a review request 7 to 14 days out. Behavior triggered, not broadcast.
Win-back campaigns fire on elapsed time relative to the customer's own cadence, not a fixed calendar. The standard trigger is 1.5 to 2 times the average gap between purchases. If a cohort reorders every 45 days, the job runs at day 65 to 70. That requires per-customer interval computation, which is exactly the kind of thing that never gets built when retention is owned by a marketing calendar.
Replenishment reminders are the same pattern applied to consumables, timed against predicted depletion rather than a monthly send.
SMS is worth wiring in for a narrow set of events, since open rates run above 90 percent against 20 to 25 percent for email. Restrict it to back-in-stock alerts, VIP windows, and loyalty balance nudges. Over-sending burns the channel permanently.
What To Build First
Instrument before you automate. If repeat purchase rate and cohort churn are not on a dashboard someone actually looks at monthly, every downstream decision is guesswork.
Then build the post-purchase sequence, because it is the cheapest and it operates during the window when the customer's opinion of the brand is still forming. Then the win-back trigger, because it recovers revenue you have already paid to acquire.
None of this is glamorous work. It is queries, scheduled jobs, and state machines keyed to order events. It also happens to be the difference between a store that compounds and one that rents its customer base from an ad platform at next year's prices. The complete guide to ecommerce customer retention covers the metrics, the loyalty mechanics, and the post-purchase systems in more depth.
What does your team track after checkout?
Top comments (0)