What Is Cohort Analysis? The Tool That Shows Which Customers Are Worth the Most

Total sales rising every month can hide a worrying truth, such as new customers who buy once and disappear forever, while the overall numbers still look good because new customers keep flowing in to replace them. Cohort analysis is the tool that reveals what the totals hide.

Cohort analysis groups customers by when they first interacted with your business, such as the month they signed up or made their first purchase, then tracks each group's (cohort's) behaviour over time, instead of looking at all customers lumped together.

For example, instead of asking how many customers repurchased this month, which mixes old and new customers and is hard to interpret, cohort analysis asks what percentage of customers who signed up in January came back to buy in months 2, 3 and 6 afterwards, which gives a much clearer picture.

Totals such as monthly sales or total active users have one major weakness: they obscure the real trends. A business with a steady stream of new customers may see total sales grow every month while, in reality, existing customers are leaving (churning) at a very high rate, masked by the new customers replacing them.

Without cohort analysis, a business has no way of knowing this problem is happening until new customers can no longer replace the ones lost, and by then it's usually too late to fix.

A standard cohort table is laid out as a matrix. Each row is a group of customers by starting month (such as the January, February and March cohorts), and each column is the number of months since the start (Month 0, Month 1, Month 2 and so on). The number in each cell is the percentage of that cohort still active in that month.

A common pattern when reading a cohort table is that numbers gradually decline from left to right in each row, as some customers drop away over time. What to watch is how steep the decline is, and whether newer cohorts have better or worse retention than older ones.

Cohort analysis isn't limited to measuring retention. It can also analyse customer lifetime value by acquisition source, for example comparing whether customers from Facebook campaigns or from Google Ads repurchase more over the long term.

This matters because a channel that seems to have a higher customer acquisition cost (CAC) may bring in better-quality customers who repurchase more over time, making it better value than a low-CAC channel whose customers buy once and disappear.

1. Define the starting point of each cohort such as the date of first sign-up or first purchase
2. Define the metric you'll track such as repurchase rate, app usage rate or cumulative purchase value
3. Group customers by starting month or week then build a table comparing cohorts
4. Find where retention drops most sharply so you know when to intervene, for example sending a promotion in month 2 if you find many customers drop off then
5. Compare cohorts by acquisition source to assess the quality of each marketing channel fairly

Cohort analysis changes the question from “What does the big picture look like?” to “Which customer groups actually create value for the business?” It helps you spot retention problems early, before the totals start to show them, and assess marketing channel quality more accurately than looking at cost per new customer alone. If you want to know which customers are worth the most, the Markettung team is ready to help analyse your cohorts and give you a clear picture.