Build the synthetic comparison for one lender's go-live
Causal inferenceHardProblem. A large lender went live with Protect at the start of a quarter, with no holdout. We want to estimate how much Protect reduced its fraud losses, using other lenders on the Plaid network that don't use Protect. How do you build the comparison, and how do you know whether the result is real?
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.