Metrics question

You noticed that Brex Credit Card activation rate went down by a lot. How would you find out 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 for a fintech activation funnel: can you break down where in the credit card activation flow users are dropping and why.

Confirm the drop is real and isolate it, checking if it is across all channels and card types or concentrated in one, like a specific acquisition partner or card tier.
Break activation into steps: card received, identity verification, first card load or setup, and first transaction, since each has a different likely failure cause.
Check for a recent change first, since a sudden activation drop often follows a change to the verification flow, a new partner integration, or a shipping or fulfillment delay.
Hypothesize by step: verification failures (a new KYC vendor issue), fulfillment delays (cards arriving late or lost), or onboarding friction (a confusing app setup step added recently).
Segment by acquisition channel, since a new partner or campaign bringing in lower intent users could also explain the drop without anything being broken.
Define the validation plan: once isolated, confirm activation rate recovers in the affected segment specifically after the fix, not just overall.

Breaks activation into discrete funnel steps rather than treating it as one number, since each step has different likely causes.
Checks for a recent change as the first, highest probability hypothesis before broader speculation.
Considers a non technical cause, lower intent users from a new channel, not just something being broken.

Investigating only technical bugs without considering channel mix or user intent changes.
Treating activation as one flat number instead of a multi step funnel with different failure points.

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