Metrics question

Imagine you are PM at Facebook. You walk into your office one day and see Messenger DAU are down by 10%. Why?

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 sudden metric drop, and disciplined segmentation before jumping to conclusions.

Confirm the drop is real, not a tracking or logging bug, before investigating further.
Segment by platform, region, and user cohort, new versus existing, to localize the drop.
Check internal causes: a recent release, an outage, or a UX change.
Check external causes: a competitor's feature launch, a seasonal dip, or a platform-level OS change.
Once localized, form a hypothesis and propose a fix or further instrumentation.
State what you'd tell your manager and by when you'd have an answer.

Rules out a measurement or logging error first, often the actual cause of sudden drops.
Segments methodically across platform, geography, and cohort instead of guessing at a single cause.
Names a plausible internal cause, e.g. a recent notification permission change reducing re-engagement.
Proposes a clear next step and timeline rather than ending on 'I'd investigate more'.

Jumping to a root cause without first checking if the data itself is correct.
Investigating everything at once instead of segmenting to isolate where the drop concentrates.

Which segment would you check first and why?
How would you tell your manager the drop is being addressed before you have the root cause?