Cohort

5 min read

Also known as: User cohort, Cohort group

A cohort is a group of users sharing a common starting event or trait, tracked over time to measure retention, conversion, or behavior patterns.

Definition

A cohort is a defined group of users who share a common characteristic or starting event — most often the week or month they signed up, made a first purchase, or entered your funnel. Instead of looking at all users in aggregate, you isolate the group and watch how it behaves over time.

Operators use cohorts to answer questions like 'Are leads from October converting better than September?' or 'Do users who came through the demo form retain longer than users from the chat widget?' Each cohort gets tracked across the same time intervals (day 1, day 7, day 30) so you can compare apples to apples.

A cohort is not the same as a segment. Segments are filters applied to your whole user base at a moment in time (e.g. 'all enterprise accounts'). Cohorts are time-anchored — they're tied to when something happened, which is what makes them useful for trend and retention analysis.

Why It Matters

Aggregate metrics lie. If your overall conversion rate is 4%, that number hides whether your newest leads are converting at 8% or 1%. Cohort analysis exposes whether your funnel is improving, decaying, or holding steady — and lets you tie changes back to the campaign, channel, or product update that caused them.

Teams that skip cohort tracking end up chasing the wrong fires. You'll celebrate a record signup month while your 30-day activation rate quietly collapses, or you'll kill a channel based on a single bad week when the cohort that came through it actually has the best LTV. Without cohorts, you're optimizing blind.

Examples in Practice

A SaaS sales team groups every lead by the week they filled out the demo form. Twelve weeks in, they see that leads from weeks 4-6 (when they ran a webinar series) closed at 22% versus 11% for other weeks — justifying a permanent webinar program.

A 30-person agency tracks client cohorts by onboarding month to measure churn. They notice clients onboarded in Q1 churn at twice the rate of Q3 clients, trace it back to a rushed kickoff process during a staffing gap, and fix the onboarding SOP.

An ecommerce operator splits buyers into cohorts based on first-purchase channel (paid social vs. organic vs. email). The paid social cohort has higher first-order value but 60-day repeat rate is half that of email — shifting budget toward email acquisition.

Frequently Asked Questions

What is a cohort and why does it matter?

A cohort is a group of users grouped by a shared event or trait — typically the date they entered your funnel — and tracked over time. It matters because aggregate metrics hide whether your business is improving or decaying. Cohort analysis lets you see if leads from this month are behaving better or worse than last month's, which is the only honest way to measure progress.

How is a cohort different from a segment?

A segment is a filter applied to your user base at one moment in time, like 'all paid users in California.' A cohort is time-anchored — it groups users by when something happened, like 'everyone who signed up in March.' Segments answer 'who are they right now?' Cohorts answer 'how is this group behaving over time compared to other groups?'

When should I use cohort analysis?

Use cohort analysis any time you need to measure retention, repeat behavior, or whether a change you made is working. It's essential for tracking funnel conversion over time, measuring the impact of product or onboarding changes, evaluating acquisition channel quality, and diagnosing churn. If you're only looking at totals or averages, you're missing the trend.

What metrics measure cohort performance?

Common cohort metrics include retention rate at day 1/7/30/90, conversion rate from signup to paid, average revenue per user over time, repeat purchase rate, and lifetime value by cohort. The pattern matters more than any single number — you're looking at how a curve shifts from one cohort to the next, not just the absolute value.

What's the typical cost of cohort analysis?

Cost depends on data volume and tooling. Basic cohort tables in a spreadsheet or BI tool cost nothing beyond analyst time. Mid-market product analytics platforms typically run a few hundred to a few thousand dollars per month depending on event volume. Most modern funnel and CRM platforms include cohort reporting as a standard feature with no separate fee.

What tools handle cohort analysis?

Cohort analysis is built into product analytics platforms, customer data platforms, BI tools, and most modern CRM and funnel software. Lighter-weight options include spreadsheet templates and email marketing platforms that report by signup date. The right tool depends on whether you need cohorts on product events, revenue, or marketing funnel stages — many operators end up using more than one.

How do I implement cohort tracking for a small team?

Start by picking one event that defines entry into the cohort — usually signup date or first purchase. Group users by week or month of that event, then track 2-3 outcome metrics across fixed intervals (day 7, 30, 90). A spreadsheet is fine for the first 90 days. Graduate to a dedicated tool once you have enough volume that manual updates become painful.

What's the biggest mistake teams make with cohorts?

The biggest mistake is comparing cohorts at unequal maturity. A cohort from last month will always look 'worse' than one from six months ago because it has had less time to convert or retain. Always compare cohorts at the same age — day 30 of the March cohort versus day 30 of the April cohort — never raw cumulative numbers.

How big does a cohort need to be to be useful?

There's no hard minimum, but cohorts under roughly 30-50 users get noisy fast — one or two outliers can swing the percentages. For small teams, widen the cohort window from weekly to monthly to increase group size. For high-volume businesses, weekly or even daily cohorts give you faster signal on whether changes are working.

Can cohorts be defined by something other than signup date?

Yes. While date-based cohorts are most common, you can cohort by any shared starting event: first purchase, first feature use, channel of acquisition, plan tier at signup, or campaign source. Behavioral cohorts (e.g. 'users who completed onboarding step 3 in week one') are especially useful for diagnosing which actions drive long-term retention.

The AMW Suite