The watch-time lift is fading over the days-in window — real or novelty?

Novelty & durationHard

Problem. A Home change showed a +6% watch-time lift on day one, but over two weeks the daily lift has been sliding. Is the effect real, and how do you report 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.

Interview tips

Strong-answer signal: naming the novelty effect, reading the stabilized plateau off the days-in curve, and deciding by whether it asymptotes above zero. Common trap: reporting the day-1 peak, or averaging the whole window and overstating the durable effect.

A/B Test & Causal Inference Questions0 / 220
← The model wins offline but loses online — which do you trust?Average view duration is a ratio — why can't you just t-test it? →