Analytics

3 articles

Cohorts, funnels and experiments: reading product data honestly, including the numbers that flatter you.

3 articles

  • Retention curves that fall steeply and then level off

    Retention curves, explained

    The single most useful chart in product analytics is also one of the most misread. How to build a retention curve, what its shape tells you and the mistakes that make it lie.

    Owen MacLeod 8 min read

  • A leaderboard of bars with a flat line running across them

    The metrics that flatter you

    Sign-ups, page views and total users all go up and to the right. That is exactly why they are so dangerous to run a product on.

    Daniel Côté 5 min read

  • Two bar charts side by side, one taller than the other

    Six ways an A/B test lies to you

    A green result on the dashboard is not the same as a true one. The most common ways experiments mislead product teams, and how to guard against each.

    Owen MacLeod 7 min read