TikTok DAU decline: validate, decompose and find the root cause

You are a product manager at TikTok and notice that daily active users (DAU) have fallen. How would you work out why? ```hint Decompose the metric DAU on any day is made up of people who were active the day before and came back, brand-new users, and lapsed users who returned. Find out which of these flows changed, and where, before reaching for explanations. ``` ### Clarifying Questions - How large is the drop, did it appear overnight or build over weeks, and is it bigger than the usual weekday and seasonal swings? - What counts as active: opening the app, or watching at least one video? Are the web version and any lighter app versions included? - Is the drop in every country and on both iOS and Android, or concentrated somewhere? - Is this viewer DAU, creator DAU, or both? ### What a Strong Answer Covers - Sizing the drop against normal variation and ruling out tracking or pipeline problems - Splitting DAU into new, retained and returning users, then segmenting by region, platform, app version, tenure and user type - Internal and external hypotheses, each paired with the data that would confirm or rule it out - Ordering hypotheses by how well they fit the observed pattern, and confirming the cause with the release timeline, holdouts or experiments - A response plan and the metrics to watch during recovery ### Follow-up Questions - If the entire drop is among users in their first week, which parts of the new-user experience would you inspect first? - DAU is down but total watch time is flat. What does that suggest, and does it change how urgent the problem is? - A competitor launched a promotion the same week you shipped a feed ranking change. How would you separate the two effects?