Cohort Analysis
Cohort analysis groups users by a shared starting event and tracks their behaviour over time. It reveals whether retention is improving or quietly getting worse.
Cohort analysis tracks groups of users who started at the same time and watches how their behaviour changes month by month.
What Cohort Analysis Means in Marketing
An aggregate retention number tells you the percentage of users still active today. Cohort analysis tells you something much more useful: whether users who joined six months ago are staying longer than users who joined a year ago.
That distinction matters. If your overall retention looks stable but your newer cohorts are churning faster, you have a problem that the headline number is hiding. If your newer cohorts are staying longer, whatever you changed recently is working. Cohort analysis makes those trends visible before they show up in revenue.
Spotify uses cohort thinking throughout their product and marketing decisions. Wrapped, their annual personalised data campaign, is partly designed to trigger a retention moment at the end of the year, re-activating users whose engagement has slipped. The event is timed to the acquisition cohort behaviour they see in their data.
How Cohort Analysis Works
The standard approach uses a retention table:
- Group users by their start date, usually the week or month they first engaged or purchased.
- Track how many from each group return in subsequent periods.
- Express each period as a percentage of the original cohort size.
The result is a grid. Rows are cohorts. Columns are the periods that follow. Each cell shows retention for that cohort at that point in time.
Retention % = Active users in period N ÷ Total users in cohort × 100
Reading down a column tells you how different cohorts compare at the same stage of their lifecycle. Reading across a row tells you how a single cohort ages.
Cohort Analysis Example
A fitness app launches a coaching feature in March. Looking at their cohort table, they notice that users acquired after March retain at 45% in month three, compared to 28% for users acquired in January. The coaching feature appears to be the differentiator. Without cohort analysis, they would have seen only a blended retention figure that obscured the improvement.
Why Cohort Analysis Matters for Marketers
You cannot improve retention if you cannot measure it accurately. Blended metrics hide whether your product is getting better or whether you are simply acquiring more new users to replace the ones leaving.
Cohort analysis also changes how you evaluate acquisition channels. A channel that brings in cheap users who churn in week two costs more than a channel that brings in expensive users who stay for a year. You cannot see that without tracking cohorts by source.
Frequently Asked Questions
What is a cohort in analytics?
A cohort is a group of users who share a common starting event within a defined time period. The most common is an acquisition cohort: everyone who signed up or made their first purchase in a given week or month. You then track how that group behaves in the weeks or months that follow.
What is the difference between cohort analysis and segmentation?
Segmentation groups users by who they are, such as location or device type. Cohort analysis groups them by when they started and tracks what they do over time. Segmentation describes your audience. Cohort analysis describes how their behaviour evolves.
How do you read a cohort retention table?
Each row is a cohort, usually a month or week. Each column is the time elapsed since their starting event. The number in each cell is the percentage still active. Healthy retention curves flatten out after an initial drop. If they keep falling to zero, you have a product problem, not a marketing problem.