Glossary · Conversion & retention
What is cohort analysis?
Cohort analysis is a method of grouping users by a shared starting event — usually the week or month they first visited or signed up — and comparing how each group behaves over the same time since that start.
Also called: Cohort, Cohort table
Updated
How does cohort analysis work?
Each row of a cohort table is a group ("visitors first seen the week of Sept 1"). Each column is time since the start (day 1, day 7, day 30). Each cell is the share of that cohort that did something — came back, paid, stayed subscribed. Reading down a column compares cohorts at the same age; reading across a row shows one cohort aging.
| First seen (week of) | Visitors | Day 1 | Day 7 | Day 14 | Day 30 |
|---|---|---|---|---|---|
| Aug 4 | 1,200 | 14% | 8% | 6% | 4% |
| Aug 11 | 1,450 | 15% | 9% | 7% | 5% |
| Aug 18 | 2,900 | 6% | 3% | 2% | 1% |
| Aug 25 | 1,300 | 16% | 10% | 8% | — |
Cohort analysis example
The Aug 18 cohort is twice the size and retains a third as well: that was the week a post went viral on a forum. Total visitors jumped, but those visitors didn't come back. The Aug 25 cohort, after a docs rewrite, retains best so far — its day-30 cell is still blank because 30 days haven't passed.
Why cohort analysis matters
Aggregate metrics mix old and new users and hide change. Cohorts isolate it: they show whether the product is getting better at keeping people, and whether a channel brings people who stick around.
Common cohort analysis traps
- Comparing incomplete cells. Recent cohorts haven't had time to reach later columns; a 0% that should be blank misleads.
- Tiny cohorts. Percentages on a few dozen users swing wildly.
- Identity resets. Cookieless or cleared-cookie visitors who return appear as new, depressing retention.
- Wrong starting event. First visit and sign-up cohorts answer different questions.
How VisitTrack does cohort analysis
VisitTrack's Retention tab groups visitors by the week they were first seen (weeks start on Monday, UTC) and shows the share who returned at day 1, 7, 14 and 30. Cells whose window hasn't fully passed show "—" instead of a misleading 0%. Retention needs visitors to be recognized across days, so it is limited in cookieless mode, and the tab says so while a site is cookieless.
Frequently asked questions
What is a cohort in analytics?
A cohort is a group of users who share a starting event in the same period, for example everyone who first visited or signed up in the same week.
What is the difference between cohort analysis and segmentation?
Segmentation splits users by an attribute such as country or plan. Cohort analysis splits them by when they started and compares groups at the same age, which makes it the standard way to measure retention.
Related terms
- Retention rateRetention rate is the percentage of users or customers from a starting group who are still active — still subscribed, or still coming back — after a given period.
- Churn rateChurn rate is the percentage of customers (or recurring revenue) that a business loses during a period, out of the customers (or revenue) it had at the start of that period.
- Activation rateActivation rate is the percentage of new sign-ups who reach a defined "activation" milestone — the first moment they get real value from the product, such as creating a first project or seeing their first data.
- Unique visitorA unique visitor is a distinct person — in practice a distinct browser or device — that visited a website at least once during a given period, counted only once no matter how many pages they viewed or how often they returned.
- Net revenue retention (NRR)Net revenue retention (NRR) is the percentage of recurring revenue a business keeps from an existing group of customers over a period, after adding expansion and subtracting downgrades and churn, excluding new customers.
Tools and guides
See which channels actually bring paying customers
VisitTrack is cookie-free analytics with revenue attribution built in. One script tag, no consent banner, live in two minutes. 14 days free, no card required.