Module
Module 5 of 6Lesson 2 of 3~17 min

Cohorts and retention, knowing whether customers stay

Retention tells you whether your product creates lasting value, yet the number shifts a lot depending on how it is calculated. This lesson teaches you to read a cohort table, judge whether usage holds over time, and compare groups without crediting a product change with an effect it did not have. Solid footing for product reviews.

Take this lesson: your first course is freeSee the full syllabus

Lesson objective

By the end of this lesson, you will be able to build and read a cohort retention table (activity definition, calculation mode, unit, period), say whether usage levels off, and compare cohorts or segments without mistaking their difference for a causal effect.

Topics covered

  • cohort analysis
  • retention curve
  • retention table
  • product retention

Where it fits

Analyse usage

Where do users drop off, who comes back, and which tool shows it?

Lessons in this module

  1. Build and read a funnel
  2. Cohorts and retention, knowing whether customers stay (this lesson)
  3. Choose your product analytics tool

What you will learn in the course

This lesson is part of the course Instrument a product and use data to make decisions

  • Build a product's metrics tree (North Star, input metrics, guardrails) from the decisions to be made, and rule out vanity metrics.
  • Design an implementable tracking plan (events, properties, naming convention, identity and accounts, emitting source, owners).
  • QA and monitor event collection to get reliable data, and know what to do when it is not.
  • Collect usage data in line with the GDPR and the CNIL rules (consent, audience measurement exemption, minimisation, retention period).
  • Analyse a conversion funnel, activation and cohort retention, and interpret the resulting curves correctly.
  • Choose a product analytics tool (PostHog, Amplitude, Mixpanel, GA4, data warehouse) based on needs, constraints and cost.
  • Turn an analysis into a reasoned decision (dashboard, decision memo) while avoiding interpretation traps (vanity, Simpson, correlation, small counts).

Related courses