Authorization rate is a ratio and value is heavy-tailed — analyze it right
Variance & ratio metricsHardProblem. Your experiment reports on authorization rate and on chargebacks. Someone runs a simple two-sample t-test on per-transaction approval and calls a 0.3pt lift "significant." And chargebacks are too rare to move. What's wrong, and how do you analyze both correctly?
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.