Cohort vs repeat rate analysis

Many startups track their repeat rate, defined as the number of customers who purchased in a specific time period divided by the aggregate number of customers who have ever purchased from the company, as a sign of the health of their business. However, this isn’t necessarily accurate.

Here’s the formula for a company’s repeat rate at time 1:

Repeat rate at time 1 = Customers who first purchased at time 0 and also purchased at time 1 / (Customers who first purchased at time 0 + Customers who first purchased at time 1)

As this formula shows, two factors contribute to a high repeat rate at time 1.

The first factor is a high percentage of customers who first purchased at time 0 also purchasing at time 1. This is the subsequent purchasing behavior of the cohort of customers who first purchased at time 0.

The second factor is a low number of customers who first purchase at time 1.

In other words, a high repeat rate can be the result of strong cohorts but it can also be the result of low growth.

You want to isolate the former effect. And this makes cohort analysis a more useful tool than the analysis of repeat rates.

Also published on Medium.