Retention measurement: defining active use, cohort windows and revenue retention

How would you define and measure retention for a product? ```hint Start from the usage rhythm Decide how often a satisfied user would naturally come back before you pick the window you measure retention over. ``` ### Clarifying Questions - Should I anchor on a specific product: a free consumer app, a weekday work tool such as Teams, or a paid subscription such as Microsoft 365, where the renewal is what matters? - Is retention measured for individual users, for paying accounts and organizations, or for revenue? - What counts as being retained: opening the product, or completing the action that delivers its value? - Is the metric for an ongoing health dashboard, or for comparing groups of users after a specific change? ### What a Strong Answer Covers - A definition of an active user tied to the product's core value rather than to logins. - Cohort-based measurement with a window that matches how often the product is naturally used. - The choice between retention variants (active in a given period, in it or any later one, or within a range) and why one fits. - Separate views for users, paying accounts and revenue where the product is sold by subscription or seat. - Data issues that distort retention numbers, and how to read cohort curves over time. ### Follow-up Questions - Week-4 retention of new users rose, but daily active users are flat. How can both be true? - How would you decide whether a retention lift in the newest cohort is real or noise? - For a product sold to organizations, how would you spot a likely cancellation before the renewal date?