Uber Win-Back Credit for Lapsed Users: Choosing the North Star and Guardrail Metrics

Uber is planning a campaign that offers dormant users a \$10 credit to get them using the app again. As the product manager who owns the campaign, pick the North Star metric you would use to judge whether it succeeded, define it precisely and explain why you chose it. Then name the supporting metrics and guardrails you would track next to it. ```hint Look past the redemption Ask what Uber wants from a returning user once the credit has been spent, and how you would tell the users the credit brought back apart from users who would have returned anyway. ``` ### Clarifying Questions - What makes a user dormant: how long without a trip or order, and does the campaign target riders, food-delivery customers or both? - What can the credit be spent on, is there a minimum fare or a cap, and does it expire? - Through which channels will users learn about the credit: push notifications, email, in-app messages, or paid ads for people who have uninstalled the app? - Is the goal a short-term lift in trips or lasting reactivation, and is there a budget or a required return on the campaign's cost? ### What a Strong Answer Covers - A campaign goal stated before any metric, separating a one-time return from a renewed habit. - One North Star metric with a precise definition, population and time window, and a reason it beats simpler candidates such as redemption rate. - Supporting metrics along the path from receiving the credit to repeat trips, including cost efficiency. - Guardrails on campaign cost, marketplace health for existing riders and drivers, and credit abuse. - A measurement design, such as a holdout group, that isolates the campaign's incremental effect. ### Follow-up Questions - Many reactivated users stop riding again a few weeks after using the credit. What do you change about the campaign or the metric? - Finance asks whether the campaign paid for itself. How would you calculate that? - How would you decide whether \$10 is the right amount, or whether different dormant segments should get different credits?