Go/No-Go Call on a Personalized Recommendations Module for the Facebook Home Page

You are a product manager at Facebook. An engineer brings you an idea: add a feature to the Facebook home page that shows each user items they might be interested in. Would you build it? Walk through how you would make the call and what you would recommend. ```hint What would it replace Pin down what the "items" are and which goal of the home page the feature would serve. Then ask what it would push off the screen: space on the home page is never free. ``` ### Clarifying Questions - What are the "items": Marketplace listings, groups, pages, events, posts from accounts the user does not follow, or a mix? - Where would the feature sit: a slot inside the feed or a fixed area of the page, and on mobile, desktop or both? - Does anything back the idea yet, such as user research, a prototype or a model the engineer has already built? - Which goal does the home page team own right now, and what planned work would this compete with for engineering time? ### What a Strong Answer Covers - Clarifying the proposal and tying it to Facebook's mission and the home page's goals before giving a verdict - The user problem the feature would solve, for whom, and how it differs from the suggestions Facebook already shows - Costs and risks: displaced friends' content and ads, recommendation quality, privacy and the feeling of being watched, integrity of recommended items, engineering effort - Success metrics with guardrails, and a staged experiment with ship and stop criteria set in advance - A clear recommendation, and how it is communicated to the engineer ### Follow-up Questions - In the test, people click the new feature often, but comments and reactions on friends' posts go down. Do you ship it? - Some users say the recommendations make them feel Facebook knows too much about them. What do you change? - If you decide not to build it, how do you tell the engineer, and what do you offer instead?