Metrics question

How would you triage a 10% drop in rider cancellations?

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

Metrics triage skill: quickly narrowing a broad metric drop to a specific, actionable cause.

Clarify direction: a drop in rider cancellations could be good, so confirm whether this means fewer cancellations or a concerning drop that hides another problem, like rides not being offered.
Check if the drop correlates with a drop in ride requests overall, since fewer requests trivially means fewer cancellations.
Segment by city, time of day, and driver supply level to localize where the change occurred.
Check for a recent product change, like a new cancellation fee or updated ETA display, around the same time.
Cross check driver side metrics, like driver cancellation rate, to see if the whole marketplace shifted together.
Propose the top hypothesis and outline a quick data pull to confirm before recommending action.

Questions whether a drop in rider cancellations is actually good or masking reduced ride requests overall.
Proposes checking for a correlated driver side metric shift, since marketplace changes often move both sides together.
Names a concrete recent-change hypothesis, like an updated cancellation fee policy, as the first thing to check.
Segments by city and time of day rather than treating the drop as one uniform, global event.

Assuming a drop in cancellations is automatically positive without checking what is really driving it.
Skipping segmentation and treating the metric as uniform across all markets.

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