Repeat Purchase Rate: The Metric That Decides What You Can Pay

The short version

Repeat purchase rate is the share of customers who buy again.

javascript

Repeat rate = customers with 2+ orders / total customers

It sets the ceiling on what you can pay to acquire someone, which makes it the most consequential number in ecommerce that most teams measure carelessly.

The two mistakes that make it useless

Measuring it blended. A blended repeat rate mixes customers acquired last week, who have had no chance to repeat, with customers from two years ago, who have had every chance. The number moves whenever your acquisition volume changes, independent of customer behavior.

Measure by cohort: of customers acquired in a given month, what share had bought again by 30, 60, 90 days.

Measuring it without a window. "Our repeat rate is 30%" is meaningless without a period. 30% within 90 days and 30% ever are very different businesses.

Segment it or it will mislead you

Repeat rate varies enormously by how the customer arrived, and averaging hides exactly the differences you need.

By acquisition source. Customers from different channels repeat at different rates. Ranking channels on first-order ROAS while ignoring this systematically overpays for the channel with the worst repeat behavior.

By first product purchased. Entry products differ in how well they lead to a second purchase. Some products are excellent first purchases and mediocre revenue; advertising them can still be correct.

By discount depth on first order. Usually the largest split. Full-price first buyers repeat differently from deep-discount buyers, and often the difference is stark enough to change channel strategy.

By cohort over time. If recent cohorts repeat worse than older ones, you are scaling into a weaker audience, and your LTV assumptions are stale.

What it lets you decide

Once measured properly, repeat rate answers:

How much you can pay to acquire. Higher repeat supports a higher first-order loss.
Which channel is actually best. Not the one with the best first-order ROAS — the one with the best customers.
Which entry product to advertise. The one that leads to a second purchase, not necessarily the one with the best margin.
Whether retention or acquisition is the constraint. If repeat rate is falling while acquisition holds, more advertising will not fix the business.

The signal worth watching

Recent cohorts repeating worse than older ones at the same age.

This is the earliest reliable warning that scaling is degrading customer quality, and it appears months before blended metrics move. It is the ecommerce equivalent of watching marginal rather than average cohorts in app UA.

Most teams do not look at it because it requires cohort tracking that ad platforms do not provide.

Practical rule

Measure repeat rate by cohort, by window, and by acquisition source, then set acquisition targets per channel from its own repeat behavior. A blended repeat rate applied to all channels overpays for some and underfunds others, every time.