Metrics question

You are a PM at Amazon and are responsible for the search feature. You have been asked to decide how many search results should be displayed on a single page when searching on Amazon. How would you go about it?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Whether the candidate can reason through a UX and performance tradeoff using data and experimentation rather than intuition alone.

Clarify the goal: balance conversion (finding the right product fast) against page load performance and cognitive overload from too many results.
Identify relevant metrics: click-through rate by result position, conversion rate, page load time, and pagination or infinite-scroll engagement.
Consider device and context differences: mobile users likely need fewer results per screen than desktop users due to screen size and load time sensitivity.
Propose testing a few options (for example 24, 48, 60 per page) via an A/B test rather than picking a number by intuition.
Define the decision metric as conversion rate and time-to-purchase, with page load time as a hard guardrail.

Frames this as an empirical question to A/B test, not something to decide from first principles alone.
Separates mobile and desktop as needing potentially different answers rather than one number for all devices.
Sets page load time as an explicit guardrail, since more results per page directly trades off against speed.

How would you decide the test is conclusive versus needs more time?
Would you personalize the results-per-page count instead of using one fixed number?