Defining a Metric Framework for Instagram's Discovery Surfaces

You are a product manager at Instagram. How would you tell whether the app's discovery features are succeeding? ```hint Walk through one discovery Trace what happens from the moment someone sees a post from an account they do not follow to the point where that find has clearly paid off for them, and ask what you could observe at each step. ``` ### Clarifying Questions - Which surfaces count as discovery here: the Explore tab, suggested posts and accounts in the feed, recommended Reels, search, or all of them? - Is this about the ongoing health of features that are already mature, or about judging a specific change or launch? - Which outcome matters most: viewers finding new interests, creators reaching people beyond their followers, or the engagement and ad revenue that discovery brings in? - Should the metrics cover all users worldwide, or concentrate on one group of users? ### What a Strong Answer Covers - A scoped definition of discovery and the job it does for viewers and for creators, tied to Instagram's goals - The path a user takes through a discovery surface, with a metric at each stage from first exposure to lasting value - One north star metric, compared with alternatives and checked for how it could be gamed - Guardrails that would catch discovery improving at the expense of the wider app or of the people using it - How the metrics would be attributed to discovery, segmented and used to judge experiments ### Follow-up Questions - After a ranking change, time spent on Explore rises but follows made from Explore fall. Would you ship the change, and what would you check first? - How would you measure what discovery does for creators, and how would you notice if the largest accounts were taking most of the exposure? - New users follow almost no one, so discovery is most of what they see. How would you read their metrics differently from those of long-time users?