Friend Suggestions vs an Ad in a Meta Feed Slot: Trade-offs and Per-User Choice

You are a product manager working on the Facebook News Feed at Meta. Two kinds of content compete for the same placement in the feed: the "People You May Know" unit, which suggests people the user might want to add as friends, and an ad. Explain the trade-offs between showing one rather than the other, and describe how you would decide which of the two a particular user should see. ### Clarifying Questions - Is this one specific placement (for example, a fixed position in each feed load), or the broader question of how often each user sees each unit? Is showing both in the same session ruled out? - Which goal matters most to the feed team right now, growing friend connections and engagement or ad revenue, and is there a revenue floor the feed has to hold? - Which users are in scope: all Facebook users, or a segment such as new accounts with few friends, and on which platforms? - Does "an ad" mean whichever ad wins the regular auction for that placement, with a known expected value, or a particular ad product? ### Part 1 — Weigh the trade-offs Compare the two units from the point of view of the user, advertisers and Meta, over the short term and the long term. Include what is given up whenever one of them takes the place of the other. ```hint When does each pay off? One unit earns its value at the moment it is shown; the other changes what the user's feed will contain for a long time afterward. Think about who benefits from each and on what timeline. ``` #### What This Part Should Cover - The value of each unit to users, advertisers and Meta, including the negative experiences each can cause. - Short-term versus long-term value, and the opportunity cost of the placement. - How the balance shifts for different kinds of users. ### Part 2 — Decide for a specific user Describe how the system should choose, for one user at one moment, between the friend-suggestion unit and an ad, and how you would confirm that your policy beats what exists today. ```hint Compare like with like Ask which facts about this user predict whether one more friend connection would change their experience, and how that benefit could be expressed so it can be weighed against an ad impression. ``` #### What This Part Should Cover - User-level signals that tip the decision. - A comparable value for both options, a decision rule, and limits on how often each unit appears. - An experiment with success metrics and guardrails, designed with the social graph's spillovers in mind. ### What a Strong Answer Covers - A goal for the feed that both units serve, stated before comparing them. - Segmentation, since the answer differs for a new user with few friends and a well-connected long-time user. - A per-user policy rather than one global winner. - Metrics and guardrails spanning connections, engagement, ad revenue and user experience. - Validation that captures long-term effects and network spillovers. ### Follow-up Questions - Your test shows that favoring the friend unit lowers ad revenue this quarter but improves long-term retention of new users. How do you make the call, and how do you bring the ads team along? - Friend requests sent from the unit go up, but the share that get accepted goes down. What does that tell you, and what would you change? - How should the policy treat a user who has dismissed the friend unit several times?