Investigating a 10% Netflix Click-Through Decline: Diagnosis and Response

Netflix's click-through rate has fallen by 10%. How would you figure out why, and what would you do next? ```hint Split the ratio Click-through rate is a ratio, so it can fall because clicks dropped or because impressions grew. Pin down which surface the metric covers, then look at the numerator and the denominator separately before reaching for explanations. ``` ### Clarifying Questions - Which click-through rate is this: clicks on title tiles on the home screen, clicks on search results, clicks on notifications and emails, or clicks on ads in the ad-supported plan? - Is the 10% a relative decline or a drop of ten percentage points, and what period is it compared with? - Did it fall suddenly or drift down gradually, and is it visible on every device and in every region? - Did anything ship around the time of the drop, such as a home-screen layout change, a new ranking model or a new app version? ### What a Strong Answer Covers - A precise definition of the metric and a split into clicks and impressions before any guessing about causes. - Data and logging checks done first, then segmentation by device, app version, region, member cohort and home-screen row. - Internal and external hypotheses checked in a sensible order, each matched to a dated event and a segment pattern. - A judgment on whether the decline matters, based on downstream measures such as plays, viewing hours and cancellations. - A response tied to the cause found, with a way to validate the fix and catch similar drops sooner. ### Follow-up Questions - Click-through fell 10% but viewing hours per member did not change. Is there still a problem, and what would you report? - The drop traces to a new recommendation model that shows a wider variety of titles. How do you decide whether to keep it? - How would you set up monitoring so the next drop like this is caught and explained within a day?