Measure a market-wide rollout with no clean control

Causal inferenceHard

Problem. A delivery change was rolled out to an entire market (or a few) — no randomized control. Leadership asks whether it worked. How do you measure it?

Before you reveal: say your answer out loud, as if you were in the real interview — get your reasoning across clearly first. There is no single correct answer: reading what the interviewer is really after and defending your own thinking is what makes an answer strong.

← Pick the unit of randomization for a delivery experimentDesign a switchback test for a real-time dispatch change →