Cohort Analysis
A method of grouping customers by their acquisition date and tracking their behavior over time. Reveals whether retention, revenue, and engagement are improving or degrading across successive groups of customers.
Why Averages Lie and Cohorts Tell the Truth
Your overall churn rate is an average. And averages hide everything interesting. A 5% monthly churn rate could mean every cohort churns evenly at 5%. Or it could mean your oldest cohorts churn at 2% and your newest cohorts churn at 12%. Same average, completely different stories. Cohort analysis shows you which one.
How to Read a Cohort Chart
The classic cohort chart is a triangle. Rows are months (when customers signed up). Columns are periods since signup. Each cell shows what percentage of that cohort’s original revenue or customers remain. You want the numbers to flatten as you move right — that means customers who survive the first few months tend to stick around.
What Good Cohorts Look Like
Strong SaaS cohorts show a steep initial drop (month 1-3), then flatten. If 100 customers sign up, you might lose 15-20 in the first 3 months, then only 2-3% per month after that. The flatter the curve after month 3, the stronger your product-market fit. If the curve never flattens, customers are not finding lasting value.
Using Cohorts to Make Decisions
Compare cohorts before and after product changes, pricing changes, or onboarding improvements. If you revamped onboarding in March, compare March-onward cohorts to January-February cohorts. If retention improved, the change worked. This is the most reliable way to measure the impact of product and GTM investments.
Frequently Asked Questions
How do you do cohort analysis for SaaS?
Group customers by sign-up month. Track each group's MRR, logo count, and engagement at month 1, 2, 3, and so on. Compare cohorts side by side. If January's cohort retains 85% at month 6 and June's retains 90%, your product is getting stickier. If it is going the other direction, something changed.
What should you track in cohort analysis?
Revenue retention (MRR from cohort over time), logo retention (customers remaining), activation rate (how quickly they reach value), expansion rate (revenue growth within cohort), and engagement metrics specific to your product. Revenue retention is the most important.
