Metrics question

Number of users on an Amazon website drops by 10%. What are the possible reasons, and how can we solve 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

Tests root-cause analysis for a moderate, broad traffic drop, structuring hypotheses before jumping to a fix.

Clarify the metric: confirm whether users means unique visitors, logged-in sessions, or something else, and over what timeframe the 10% drop occurred.
Rule out measurement issues: check for a tracking change, a tagging bug, or a bot-traffic filter update that could mechanically shift the count.
Segment the drop: by traffic source, device, and region, since a broad drop concentrated in one channel points to a specific cause.
Check external causes: an SEO ranking change, a paid marketing budget cut, a competitor promotion, or seasonal dip versus the prior comparable period.
Check internal causes: a recent site redesign, a checkout or login flow bug, or a performance regression increasing bounce.
Prioritize fast checks first, channel segmentation and a year-over-year comparison, then propose a fix path per confirmed cause.

Checks measurement validity before assuming real behavior changed, a step many candidates skip.
Segments by channel, device, and region before hypothesizing a cause, avoiding a one-size explanation.
Separates external causes from internal ones for a structured investigation.
Proposes fast, cheap checks first, channel segmentation and year-over-year comparison, before slower diagnostics.

Jumping to a single guessed cause without segmenting the drop by channel or region first.
Ignoring measurement or tracking issues as a possible explanation for the drop.

Chapters of the AI PM course that teach what this question tests.