Scorecard for Video Posts in the Facebook Feed: Primary, Funnel and Guardrail Metrics

You are the product manager for video in the Facebook News Feed, the ranked stream of posts people see when they open Facebook. Video is one of several formats that compete for space in that stream, next to text, photos, links and shared posts. Define the metrics you would use to decide whether video in the feed is succeeding: the one metric you would hold the product to, the supporting metrics that explain why it moves, and the guardrails that catch a gain for video that hurts the feed or the people using it. ```hint Watched versus played Feed video often starts playing on its own as people scroll. Think about how you would separate a video someone chose to watch from one that merely played on screen, and what more video could cost the rest of the feed. ``` ### Clarifying Questions - Is video in the feed already live and the task ongoing health tracking, or are we judging a specific change such as a new player or a ranking update? - Does the scope cover only videos that play inline in the feed, or also the dedicated video destination people reach by tapping a feed video? - Should the metrics cover only viewers, or also the friends, creators and Pages who post video to the feed? - Are both the mobile apps and the website in scope, given that autoplay and sound behave differently on each? ### What a Strong Answer Covers - A goal for video that follows from what the feed is for, and one primary metric that reflects intentional viewing rather than raw plays. - A funnel of supporting metrics from exposure to start, qualified view, completion and interaction, each with a precise definition. - Guardrails for overall feed health: total time and sessions per person, engagement with non-video posts, and negative feedback on video. - A separate technical layer for playback quality, such as start failures, time to first frame and rebuffering. - Leading versus lagging signals, segmentation, and the decision each metric would trigger. ### Follow-up Questions - Video watch time is up 10 percent, but total time on Facebook per person is flat. Is that a success? - How would you set the threshold for counting a view of a muted autoplay video, and what goes wrong if it is too low or too high? - A ranking change doubles video starts, but hides and "see less" taps on video posts also climb. What do you do?