Diagnose a 10% Drop in Facebook Page Creation, Then a 5% WAU Decline After the Fix

Facebook Pages are the public presences that businesses, brands, public figures and organizations run on Facebook. As a product manager, you are handed two metric problems in sequence: first a fall in the number of new Pages being created, then a second decline that shows up once the first is resolved. ### Clarifying Questions - Are both declines measured the same way, for example week over week, and did each appear suddenly or build up gradually? - Were there known launches, outages or holidays in the period, or should you discover them yourself? ### Part 1 — Page creation falls by 10% The number of new Facebook Pages being created is down 10%. How would you work out why, and what would you do about it? ```hint Localize before you explain Before listing possible causes, find out whether fewer people are starting to create a Page or fewer are finishing, and whether every kind of Page and every platform is affected equally. ``` #### Clarifying Questions for this Part - Does the count include every Page created, or only Pages whose setup was completed and that were not later removed as spam? #### What This Part Should Cover - Checking the metric's definition, comparison window and data before treating the drop as real - Segmenting the drop and walking the creation funnel to localize it - Internal and external hypotheses, ranked, each with the data that would confirm or rule it out - Whether the drop actually harms the Pages ecosystem, and the action each root cause calls for ### Part 2 — Weekly active users fall by 5% after the fix Assume you found the cause and Page creation is back to its usual level. Now weekly active users (WAU) are down 5%. What would you do? ```hint Use the timing, but test it The decline arrived after Page creation recovered. Consider what changed when it did, but treat any link between the two as a hypothesis to test rather than a conclusion. ``` #### Clarifying Questions for this Part - Which WAU fell: weekly active users of Facebook overall, people who engage with Pages, or Page admins? #### What This Part Should Cover - Defining which WAU fell and breaking the decline down by new, retained and returning users and by segment - Testing whether the Page-creation fix caused the decline, with a method that separates cause from coincidence - If the two are linked, weighing Page creation against active users and choosing a path ### What a Strong Answer Covers - An ordered diagnosis method applied consistently to both metrics, rather than a list of guesses - Prioritization by impact, including how a 5% fall in active users compares with a 10% fall in Page creation - Specific data checks, each tied to the decision its result would lead to - Communication with stakeholders and guardrails while the investigation runs ### Follow-up Questions - The fix was launched as an experiment, and the experiment shows no WAU difference between test and control. Where do you look next? - Page creation has recovered, but new Pages now post much less in their first week than before. Does that change your view of the fix? - Leadership asks whether to roll back the Page-creation fix today, before the investigation is finished. What do you recommend?