Why Cohort Analysis Beats a Single Repeat Rate Number
Posted by Admin on 7/31/2026
Most D2C brands track a single repeat purchase rate and treat it as a health metric. The problem is that one number is an average across every customer you have ever acquired, and averages hide direction. Your repeat rate can look stable while your business is quietly deteriorating underneath it.
What a single repeat rate conceals. Imagine your overall repeat rate has held at thirty percent for a year. Reassuring, until you split it by acquisition month. It turns out your customers from twelve months ago repeat at forty five percent, while customers acquired in the last three months repeat at fifteen percent. The loyal older cohorts are propping up the average while your recent acquisition quality is collapsing. The single number told you everything was fine. It was not.
What cohort analysis does. It groups customers by the month they were acquired and tracks each group separately over time. Instead of one blended figure, you see whether each successive cohort is behaving better or worse than the ones before it. That direction is the actual signal. Improving cohorts mean your product, targeting and experience are getting better. Declining cohorts mean something upstream is breaking, often that you scaled acquisition into lower quality audiences.
The refinement that matters most in early stage. Track repeat purchase from customers with no personal connection to you specifically. Early orders from your own network, friends and community will repeat out of goodwill, and they inflate your cohort numbers with a loyalty that does not generalise to strangers. Product market fit shows up when people who do not know you buy again. That is the cohort worth watching.
The takeaway. A flat average is not stability, it is a lack of resolution. Break your repeat rate into monthly cohorts and read the trend across them, because the direction of your newest cohorts predicts your future far better than the average of all of them.
Have you split your repeat rate into acquisition cohorts, and are your newest cohorts trending up or down?
