Metrics question

There's a 50% drop in the number of Uber rides per day. Diagnose the problem.

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

Structured, hypothesis-driven diagnosis of a severe, sudden metric drop.

Confirm the 50% drop is real: check for an outage, a data pipeline break, or a tracking bug before treating it as a true demand collapse.
Check timing precisely: was this instant (suggests an outage or app crash) or gradual over days (suggests a real behavior shift).
Segment by geography and platform to see if the drop is global or localized, which points to different causes (app bug versus regulation or competitor).
Check supply side: are drivers unavailable due to an app issue, driver app outage, or a mass driver deactivation event.
Check external factors: a local event, weather, a new regulation, or a PR crisis affecting rider trust.
Prioritize the most severe, most likely hypothesis first given a 50% drop is extreme and likely has one dominant cause.

Treats a 50% drop as almost certainly a systemic failure (outage, pricing bug, app crash) rather than gradual demand erosion, and investigates accordingly.
Checks both rider-side and driver-side apps, since a supply collapse looks identical to a demand collapse in raw ride counts.
Gives a concrete example, e.g. a payment processing outage blocking ride completion, as a plausible first hypothesis to rule out.
Proposes an immediate mitigation (status page, refunds, PR statement) in parallel with root-causing, given the severity.

Treating a 50% drop the same as a small gradual decline, missing that its scale implies a systemic failure.
Only investigating demand-side causes and ignoring supply-side or technical failures.

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