Cohort analysis

Cohort analysis groups customers by a shared starting point, such as the month of their first order, and follows each group over time to compare what they do.

By Berend Vrakking, founder of Product Metrics. Updated 7 October 2026.

Example

Repeat orders by first-order month
First-order month	Customers	Ordered again by month 1	By month 2	By month 3
January	200	20 (10%)	30 (15%)	36 (18%)
February	250	25 (10%)	40 (16%)	50 (20%)
March	300	45 (15%)	60 (20%)	–

Each row is a cohort, and each cell is the customers who have ordered again so far, as a share of that cohort. A dash means March has not had its third month yet. March customers came back faster than January customers. Illustrative data.

For one product, and for an account

A cohort can start from anything you can date or name. Group customers by the first product they bought and you can see which products bring people back. One account-wide repeat figure blends the product that earns loyal customers with the one that earns a single visit.

You build cohorts from your own order data. On the Google Ads side, Full Signal Tracking keeps the first order apart from the ones after it: new customers and returning customers arrive as separate conversion actions.

Common mistake

Comparing a cohort that started last month with one that started last year. The young one has barely had time to come back, so line cohorts up by months since the first order.

Questions

What does cohort mean in simple terms?

In ecommerce, a cohort is usually the customers who placed their first order in the same month. More generally, it is any group of people who share a starting point.

What is an example of a cohort in ecommerce?

All customers whose first order was in January form the January cohort. Customers whose first purchase was one particular product form a cohort too. Either way, you follow the group month by month and count how many of them order again.

How is cohort analysis different from a repeat purchase rate?

A repeat purchase rate is one number for all customers in one period. Cohort analysis keeps each starting group apart, so you can see whether newer customers order again faster or slower than older ones.

Keep reading

Repeat purchase rateThe single figure a cohort table breaks down by starting month.
LTV:CAC ratio for ecommerceRepeat orders by cohort show what a customer brings in after the first order.
Product SegmentationSee each product in one of six segments, by volume and return.
Customer acquisition cost (CAC)What each cohort cost to win.