Skip to content
Back to ProductAll lessons

Product

Retention & Cohorts

The idea, in plain English

Retention measures whether people keep coming back after they first show up. A cohort is just a group who started at the same time — like everyone who joined your gym in January. Instead of one blurry average, you follow that January group week by week: how many are still active after one week, two weeks, a month? A product can look healthy because new people keep arriving, while secretly the people who join keep leaving. Cohort retention exposes that leak.

How it works

  1. 1Pick a cohort: everyone who joined in the same period (say, week 0).
  2. 2Each later week, count how many of that SAME original group came back and were active.
  3. 3Divide each week's returners by the cohort's starting size to get a retention percentage.
  4. 4Watch the curve. A gentle flattening is healthy; a cliff means people try it once and don't return.

When you'd use it

Use cohort retention to judge whether your product delivers lasting value, not just first-time curiosity. It's the truth serum behind growth: if retention is leaking, pouring in more new users just fills a bucket with a hole in it.

Common beginner mistakes

  • Hiding behind total-user growth. Rising totals can mask terrible retention when acquisition is simply outrunning churn for now.
  • Comparing cohorts that aren't comparable. A holiday-signup cohort may behave very differently from a normal week — compare like with like.

Try it — edit and run

Click the code to edit · press ⌘/Ctrl+↵ to run

Editable code. Tab and Shift+Tab indent. Press Escape, then Tab, to move focus out of the editor.

Expected output — hit Run to try it
Cohort started with 200 users
Week 0: 200 active (100% retained)
Week 1: 120 active (60% retained)
Week 2: 90 active (45% retained)
Week 3: 72 active (36% retained)
Lost 18 users from week 2 to week 3

Not sure this is the right topic? See the learning paths → or where this leads →