Root-Cause Analysis of a Declining Netflix Viewing Completion Rate

You are a product manager at Netflix, and the completion rate for titles that members start watching is going down. Explain how you would work out why it is falling and how you would address it. ```hint Split the ratio before chasing causes Completion compares finished plays with started plays, so check which side moved, and whether the mix of titles, members and devices behind the number changed, before concluding that people now watch differently. ``` ### Clarifying Questions - How is completion defined here: reaching the start of the end credits, watching a fixed share of the runtime, or finishing a whole season? - How large is the drop, since when, and did it happen suddenly or gradually? - Is it visible everywhere, or only in certain regions, devices, plans or kinds of content? - Have other engagement numbers, such as hours watched, plays started and cancellations, moved at the same time? ### What a Strong Answer Covers - A precise metric definition and checks that rule out logging, pipeline and definition artifacts - A decomposition of the ratio and a segmentation that separates mix shifts from real changes in viewing behavior - Hypotheses across product changes, playback quality, content and external factors, ranked and tied to the data that would confirm each one - A judgment on whether the decline actually harms members or the business, using retention and viewing hours - Fixes matched to the confirmed cause, with a way to verify that completion recovers ### Follow-up Questions - If completion fell only for newly released series while total viewing hours rose, would you act, and how? - Using playback logs alone, how would you separate members who quit because of streaming problems from members who quit because they disliked the show? - Suppose a recommendation change raised plays and hours but lowered completion. Would you roll it back?