Metrics question

Set goals for Facebook notifications.

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

Tests goal-setting for a feature that can easily be over-optimized (notification volume) at the expense of user trust.

Clarify the real objective: notifications should drive re-engagement without becoming spam that increases opt-outs, so define both a growth and a guardrail goal.
Set a primary goal: lift in click-through rate on notifications that leads to a session within 10 minutes.
Set a guardrail goal: keep notification opt-out and mute rate flat or declining, since raw send volume is an anti-goal.
Segment goals by notification type, social (tags, comments) versus algorithmic (suggested content), since their healthy click rates differ.
Propose a target, e.g. improve session-driving CTR by X% quarter over quarter while opt-out rate stays under a fixed ceiling, as the combined goal.

Frames notifications as a scarce trust resource, not a free growth lever, which is the key insight interviewers want on this question.
Proposes relevance-weighted goals, e.g. a comment on your post should have a much higher CTR bar than a re-engagement nudge.
Gives an illustrative number, e.g. assume current CTR is 4% and opt-out rate is 2% monthly, and target improving CTR without moving opt-out past 2.5%.
Notes Spotify's own notification model (personalized nudges) as a comparable case of balancing relevance against fatigue.

Setting 'send more notifications' or 'increase DAU' as the goal without a guardrail against fatigue.
Treating all notification types as one bucket instead of segmenting by type and relevance.

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