Cohort Analysis: How to Track and Compare Groups of Visitors Over Time
Standard analytics shows you what happened today or this month. Cohort analysis shows you whether things are getting better or worse over time — by comparing groups of visitors who arrived in different periods and watching what they do next. It is the single most powerful tool for understanding whether your site is genuinely improving at retaining and engaging its audience.
What Is a Cohort?
A cohort is a group of visitors who share a common starting point — most often, the week or month they first visited your site. Cohort analysis tracks each group forward from that starting point and compares how the groups behave relative to each other.
For example, visitors who first arrived in January form the "January cohort." Cohort analysis lets you see what percentage of them came back in February, March, and April — then compare those numbers to the February cohort's return rates at the same intervals. If the February cohort returns at higher rates than January, something improved. If the rates are falling, something broke or the audience quality changed.
The Cohort Retention Grid
The standard output of a cohort analysis is a grid. Rows are cohorts (Week 1, Week 2, etc.). Columns are time intervals since first visit (Day 1, Day 7, Day 30, etc.). Each cell shows what percentage of that cohort was still active at that interval.
| Cohort | Day 0 | Day 7 | Day 30 | Day 60 |
|---|---|---|---|---|
| April cohort | 100% | 28% | 14% | 9% |
| May cohort | 100% | 31% | 17% | 11% |
| June cohort | 100% | 35% | — | — |
Reading down each column shows whether retention at that interval is improving across cohorts. In the example above, Day 7 retention improved from 28% to 35% — a clear positive trend. Day 30 improved from 14% to 17%, suggesting whatever you changed in May had a lasting effect, not just a novelty bump.
What Cohort Analysis Reveals That Simple Metrics Hide
Aggregate visitor counts can grow even while your site gets worse at keeping people. If you acquire enough new visitors, the total number masks a collapsing retention rate until the acquisition slows. Cohort analysis removes this illusion by locking each group to its own starting point.
- Did a product change improve or hurt retention? Compare the cohort that arrived before the change to the one after.
- Is a traffic source sending quality visitors? Segment cohorts by source and see whether organic visitors retain better than paid ones.
- Are newer visitors stickier than old ones? A rising diagonal in the grid means yes — your product is getting better at keeping people engaged over time.
- Where exactly do visitors drop off? A sharp drop at Day 7 tells you the first-week experience is the weak point, not the initial landing.
Acquisition Cohorts vs Behavioural Cohorts
The most common type is an acquisition cohort: grouped by the period visitors first arrived. A behavioural cohort groups visitors by an action they took — for example, everyone who completed a sign-up form, regardless of when they first visited. Behavioural cohorts answer questions like "do visitors who download our guide retain better than those who don't?" They are harder to set up but extremely valuable for validating content or feature investments.
How to Use Cohort Insights in Practice
- Run a cohort comparison before and after any significant content change, redesign, or new feature launch. The cohort after the change is your test group; the cohort before is your control.
- If Day 1 retention is low across all cohorts, the problem is the landing experience — visitors arrive but don't find a reason to come back. Fix the onboarding or value proposition.
- If Day 30 and beyond retention is low but early retention is fine, the problem is long-term engagement — you hook people initially but don't give them a reason to stay. Add depth: email newsletters, new content cadence, or gated features.
- Compare cohorts from different traffic sources using UTM parameters. High-volume channels that produce poor cohort retention are costing you more than they return.
Track cohort retention automatically with statpx
statpx tracks visitor return rates over time so you can spot whether your site is improving at keeping people. Privacy-friendly, no cookies required.
Start for free →The Bottom Line
Cohort analysis turns your raw visitor numbers into a longitudinal story. It answers the question every site owner should be asking: are the people who come back actually increasing over time, or is growth an illusion powered by ever-growing acquisition? Once you start reading cohort grids, you stop celebrating total traffic milestones and start caring about the only thing that compounds — visitors who return. Pair cohort analysis with your retention metrics and new vs returning visitor data for a complete picture of audience health.