Facebook Events success metrics and ten small events vs. one large event

You are the product manager for Facebook Events, the part of Facebook where people create events, invite others, discover gatherings and RSVP. The interviewer asks two connected questions: what success looks like for the product, and how you would judge between two very different outcomes. ### Clarifying Questions - Does "attendee" mean someone who tapped Going, or someone who actually showed up, and can Facebook observe the difference? - Are the events in question public events open to anyone, or private gatherings among friends and group members? - Is there a company goal Events should serve right now, such as engagement among existing users, reaching new or lapsed users, or revenue from ticketing and promoted events? ### Part 1 — Set the success criteria Define success for Facebook Events: what the product is for, the one metric that best shows it is succeeding, and the supporting and guardrail metrics you would track alongside it. ```hint Success means people actually meet Think about what a good event looks like for the person hosting it and for a guest, and look for a metric that captures both. An RSVP costs nothing to give, so ask what happens after it. ``` #### What This Part Should Cover - The purpose of Events and the two sides it serves, hosts and guests - A primary metric tied to people getting together, with its measurement limits - Supporting metrics for hosting, discovery and attendance, plus guardrails ### Part 2 — Ten 100-person events or one 1,000-person event Facebook Events could end up with either of two outcomes: 10 events with 100 attendees each, or 1 event with 1,000 attendees. How would you assess which outcome is better for the product? ```hint Same total, different shape Both outcomes add up to the same number of attendees, so the headline count cannot decide it. List what else differs between the two outcomes and judge each difference against your success criteria from Part 1. ``` #### What This Part Should Cover - Why total attendance ties, and which second-order differences decide the comparison - A recommendation, with the conditions under which it would flip - How you would settle the question with data ### What a Strong Answer Covers - One consistent set of success criteria applied in both parts - Honesty about measurement: RSVPs versus real attendance, and the privacy limits on detecting who showed up - A way to settle the comparison with evidence (a historical cohort comparison or an experiment) and the confounders to control for - Product consequences of the answer, such as how Events discovery should rank small and large events ### Follow-up Questions - Your primary metric rises, but the share of hosts who host again falls. What could explain it, and what would you do? - How would you design an experiment that shifts Events recommendations toward small local events, and what results would you need to ship it? - Would your answer change in a market where people mainly use Facebook to follow public pages and creators rather than friends? - Suppose your team's current mandate is growth, and the North Star has to serve that goal. Which metric would you make the North Star, how would you defend it, and does your answer to Part 2 change?