METRICS & ITERATION · LESSON 06.04intermediate

Funnel + cohort analysis.

Where users drop off, and how long they stay once they don't.

↳ tl;dr

Funnel analysis shows where users fall out of a multi-step flow. Cohort analysis groups users by start week and shows how each group's behavior evolves over time. Together they answer two different questions: where am I losing people, and are my fixes actually working over time.

Funnel analysis

Define the steps (visit → signup → activation → first action → recurring use). Count users at each step. Look at conversion rates between steps. The biggest drop is your bottleneck. Fix the bottleneck before optimizing anywhere else — gains elsewhere are washed out by the leak.

Cohort analysis

Group users by when they joined (the "cohort"). Track each cohort's retention over weeks/months. Cohorts that retain better than older ones = your improvements are working. Cohorts that retain worse = something regressed. Aggregate metrics hide both.

in the wild

A team celebrated rising overall MAU for six months. Cohort analysis revealed: each new cohort retained worse than the last. They were growing because of marketing spend, not product fit. Cohorts surfaced what aggregates hid.

Aha moment / activation

The action that predicts long-term retention. Facebook's famous one: 7 friends in 10 days. Slack's: 2,000 messages in a workspace. Find yours by looking at retained vs. churned users — what did the retained ones do in week 1 that the churned didn't? That's your activation event.

// sources

Sources cited

  1. [01]
    Lean Analytics: Use Data to Build a Better Startup Faster

    Croll, A. & Yoskovitz, B. · O'Reilly Media · 2013 · retrieved 2026-05

    Defines actionable vs vanity metrics and the One Metric That Matters.

// sources

Further reading

  1. [01]
    Lean Analytics: Use Data to Build a Better Startup Faster

    Croll, A. & Yoskovitz, B. · O'Reilly Media · 2013 · retrieved 2026-05

    Defines actionable vs vanity metrics and the One Metric That Matters.