Root-causing a 20% decline in new Facebook Group creation

You are the product manager for Facebook Groups. The number of new groups being created has dropped by 20%. Walk through what you would do. ```hint Real, then where Before naming causes, confirm the drop is genuine and find which users, surfaces or steps of the creation flow account for it. ``` ### Clarifying Questions - Over what window did the 20% drop happen, against what baseline (the prior week, the prior quarter, the same period last year), and was it sudden or gradual? - Is the metric the number of groups created or the number of people creating groups, and is it counted when someone taps Create or when a group is fully set up? - Is the decline spread evenly, or concentrated in one platform (app versus desktop web), region or type of group? - Have related metrics moved too, such as sign-ups, daily active users or activity inside existing groups? ### What a Strong Answer Covers - Pinning down the metric's definition, window and baseline before forming hypotheses. - Ruling out logging, pipeline and definition changes first. - Breaking the metric into parts (people who could create, people who start, completion rate, groups per creator) and segmenting by platform, region, user tenure and group type. - Internal and external causes, ordered by likelihood and cost to check, each paired with the data that would confirm or rule it out. - Judging whether the drop is actually harmful (for example, fewer spam groups) and the action that follows once the cause is found. ### Follow-up Questions - The entire drop is in groups that never gain a second member. Does that change how worried you are and what you do next? - An experiment that rearranged navigation items explains half the drop, but it improves another team's key metric. How do you resolve it? - How would you monitor group creation so that a drop like this is caught within days?