Metrics question

You're the PM of Uber. The ride rating is 20% less than the industry standards. How do you solve this?

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 diagnosis of a lagging quality metric relative to competitors and a structured path to a fix.

Confirm the benchmark: what counts as industry standard and whether the comparison is fair, same markets and same rating scale.
Break rating down into drivers: driver behavior, car condition, route efficiency, and app reliability including cancellations and ETA accuracy.
Segment the gap by city and driver tenure to see if it is broad or concentrated in specific markets.
Compare against Lyft in the same markets if data exists, to isolate whether this is Uber-specific or market-wide.
Prioritize the segment showing the largest gap, since that is where the fix has the most leverage.
Propose a fix, such as targeted driver coaching, and track the rating trend over 8 weeks to validate it.

Questions the benchmark's validity before accepting the 20 percent gap at face value.
Decomposes rating into concrete sub-drivers instead of treating it as one opaque number.
Uses segmentation by city and tenure to localize the problem before proposing a fix.
Proposes a time-bound tracking metric to validate the fix actually worked.

Accepting the 20 percent gap without questioning whether the comparison is apples to apples.
Proposing a blanket driver bonus without diagnosing the actual cause.
Not defining how you would know the fix worked.

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