Cohort analysis
Last updated 2026-06-14
Definition
Cohort analysis groups users by something they share — usually the week or month they signed up — then tracks how each group behaves over time. It separates whether newer users stick around better than older ones, instead of blending everyone into one average. Quri reads your first-party events so you can compare cohorts without exporting raw tables.
How to do this in Quri
- Connect a source or use Quri Tag so signup and activity events are captured.
- Group users by their start week or month to form cohorts.
- Track each cohort’s retention or revenue across the weeks that follow.
- Ask Quri whether a recent cohort is behaving better or worse than earlier ones.
Frequently asked
- Why use cohorts instead of a single average?
- An overall average hides change. If new users churn faster, a blended number stays flat while the trend rots. Cohorts split groups by start date so you catch the shift early.
- What can I group a cohort by?
- Most often the signup period, but any shared trait works — acquisition channel, plan, or first action. Quri builds cohorts from the first-party events it reads.