Cohort

A cohort is a group of users who share the same starting event in the same time window, such as everyone who signed up in March or started a trial in week 12. You follow that group over time to see how it retains, activates or converts, and you compare it with other groups that started earlier or later.

How it is measured

Pick a start event, such as signup or trial start, and a time bucket, such as a week or a month. Everyone who hit the start event in that bucket is one cohort. Then track one outcome by tenure.

Cohort rate at month N = users in the cohort who reached the outcome by month N, divided by all users who entered the cohort.

Put cohorts side by side as rows and tenure as columns. Read across a row to see one group mature. Read down a column to see whether newer groups do better at the same age.

A concrete case

For free-to-paid conversion, Kissmetrics describes a cohort view that "groups users by the month they signed up and follows their cumulative upgrade rate over time" (Kissmetrics). That view shows whether conversion is improving or whether you are only converting an old backlog of free users.

Mixpanel suggests splitting users into activated and non-activated cohorts. It warns: "If the difference is small, your activation definition may be wrong" (Mixpanel). The same post cites weekly retention for business-to-business (B2B) products ranging from 44.6% to 77.9% globally (Mixpanel).

Baremetrics reports that SaaS companies that routinely run cohort analysis see roughly an 18% higher conversion rate than those that skip it (Baremetrics).

Usual mistakes

• Reporting one blended monthly rate. It hides whether new signups convert better or worse.
• Treating recent cohorts as final. Their rates still move as late converters arrive (Baremetrics).
• Comparing cohorts at different ages. Compare month 2 with month 2, never month 2 with month 6.
• Changing the start event between periods. The rows stop being comparable.

When it matters

Use cohorts any time you change onboarding, pricing or acquisition channels and need to know whether the change worked. This week, build one table: signup month as rows, upgrade or retention rate by month of tenure as columns. Check it against our free-to-paid conversion benchmarks and activation rate benchmarks.