Measuring the Effect of Short-Form Video Apps on Netflix Viewing Engagement

As a product manager at Netflix, you are asked how you would assess the impact that TikTok and YouTube Shorts have on Netflix engagement. Walk through what you would treat as engagement for this analysis, the ways short-form video could plausibly change how members use Netflix, the data and methods you would use to separate that effect from everything else that moves engagement, and how you would state the result. ```hint Mechanisms before data Before choosing any data source, list the ways short-form video could change Netflix viewing, in both directions, and which members and devices each mechanism would hit first. Each mechanism predicts a different pattern, and those predictions tell you which comparisons can isolate the effect. ``` ### Clarifying Questions - Which engagement measure matters most here: viewing hours per member, days active per month, new title starts, or cancellations, and on which devices? - Is this a one-time sizing of the competitive threat, or a way to keep tracking it as these apps evolve? - Should the assessment be global, or focus on markets where short-form video adoption differs sharply, including any where one of these apps is unavailable? - Beyond Netflix's own viewing logs, can I use third-party time-use or app-usage panels and commission member surveys? ### What a Strong Answer Covers - A precise engagement definition, broken down by device, member age band, tenure, market and content type, so an effect has somewhere to show up - Plausible mechanisms in both directions, such as time displacement on phones versus discovery of Netflix titles through clips, each with a testable prediction - A credible way to isolate the effect from confounders such as the content slate, seasonality, price changes and account-sharing rules, using natural experiments, cross-market comparisons and cohorts rather than raw correlation - Triangulation across internal data, external time-use data and member research, with honesty about what cannot be measured - A result stated as a direction and size with uncertainty, by segment, linked to the decisions it should inform ### Follow-up Questions - Heavy short-form users watch less Netflix on their phones but the same amount on TVs. What do you conclude? - How would you test whether clips of Netflix shows circulating on these apps bring members back to watch? - If the effect is concentrated among the youngest members, how should that change Netflix's product priorities?