You work on a ride-sharing and delivery marketplace where a new Uber Eats feature could improve customer conversion, but it may also shift demand in a way that hurts marketplace balance. The team wants to test it online and is worried about side effects on the broader system.
How would you design the experiment so you can measure the feature’s impact without harming marketplace balance? What would you treat as the primary metric, what guardrails would you set, and how would you think about randomization and analysis?
You work on a ride-sharing and delivery marketplace where a new Uber Eats feature could improve customer conversion, but it may also shift demand in a way that hurts marketplace balance. The team wants to test it online and is worried about side effects on the broader system.
How would you design the experiment so you can measure the feature’s impact without harming marketplace balance? What would you treat as the primary metric, what guardrails would you set, and how would you think about randomization and analysis?
