Diagnosing a Sudden 50% Rise in Average Comments on a News Platform

You are a product manager for a news platform where readers can comment on stories. Over a span of two days, the platform's average number of comments rose from 2 to 3. What hypotheses would you form to explain the jump, and how would you work out which one is correct? ```hint Start with the ratio Before looking for reasons people would comment more, ask exactly how the average is calculated and whether either side of it could have moved for reasons unrelated to reader behavior. ``` ### Clarifying Questions - Is the average measured per article, per active reader, or per reader who comments? - Is the increase concentrated in particular regions, languages, news sections or app versions, or spread evenly across them? - Does the team see more comments as good, or is there concern about comment quality, such as spam or abusive threads? - Did anything ship in the days before the jump, such as an app release, a ranking change or a moderation change? ### What a Strong Answer Covers - Sizing the change and confirming it is real: the metric's definition, both parts of the ratio, and logging or pipeline changes - A structured set of hypotheses spanning measurement, internal product changes, external events, shifts in who is commenting, and abuse - Segment cuts that isolate the cause, checked in a sensible order - The specific evidence that would confirm or rule out each leading hypothesis - A judgment on whether the increase is healthy, and the action that follows from each finding ### Follow-up Questions - Most of the extra comments come from a small number of recently created accounts. What do you check next, and what do you do? - The rise traces to a ranking change that promotes more divisive stories. Would you keep the change? - How would you tell a lasting shift from a short spike tied to one news event?