Your order-rate experiment is underpowered — how do you fix it?

Variance & ratio metricsMedium

Problem. An experiment on order rate can't reach significance in a reasonable time — the metric is too noisy. How do you get power without just waiting for a huge sample?

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.