A notifications policy shipped platform-wide. Measure it without an A/B.
Causal inferenceHardProblem. A new notifications policy (fewer, higher-quality notifications) shipped to all members in one country at once, with no control group. Engagement in that country changed afterward. Give the method to estimate the causal effect and how you would validate it. (The business framing of this decision lives in the Product Case section; here, the mechanics.)
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.