Metrics question

You're the PM for Facebook Like. You notice that Like acceptance is down 5%. What do you do?

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

Root-cause investigation for a specific, unusual metric requiring clarification and structured diagnosis.

Clarify what acceptance means here, likely the rate at which a Like action completes successfully versus errors out or gets undone.
Quantify the drop across platforms and regions to see if it is global or localized to a specific client.
Check for a recent client or backend deploy affecting the button's tap handling or API call success rate.
Check for a UI change, such as a redesign that made the button harder to tap accurately, increasing accidental undos.
Segment by user type, new versus experienced users, since a UI change might disproportionately confuse one group.
Propose a fix based on the diagnosed cause and track the metric's recovery over the following week.

Explicitly clarifies the unusual metric term before diagnosing, showing good judgment on an ambiguous prompt.
Segments by platform and region first, a disciplined way to localize a technical versus behavioral cause.
Considers a specific, plausible UI cause rather than jumping to abstract explanations.
Proposes a concrete validation step to close the loop after the fix.

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