Product design question

How would you double active users on Facebook Shop weekly when only 6% of FB users remain active in the first week and 0.4% remain active after 3 weeks?

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 growth and retention problem-solving given specific, steep drop-off numbers, requiring a diagnosis-then-fix approach, not just tactics.

Restate the funnel: 6% week-1 retention and 0.4% week-3 retention means nearly all early users churn almost immediately, so the fix must target week one first.
Hypothesize why: likely weak onboarding (no clear first action), lack of relevant inventory for the user's interests, or unclear value versus browsing Marketplace directly.
Prioritize week-1 activation: a guided first-purchase or first-listing flow that gets a user to a meaningful action (saved item, first message to a seller) within the first session.
Address ongoing relevance: personalized re-engagement notifications when new items matching saved interests appear, to pull users back after day one.
Define success: shift the week-1 retention curve first, e.g. from 6% to 12%, before worrying about week-3, since week-3 retention is downstream of week-1 activation.

Recognizes the steep drop implies an activation problem, not a long-term engagement problem, so focuses the fix on session one.
Sequences the fix logically, fix week-1 activation before expecting week-3 retention to move, rather than attacking both simultaneously.
Uses the given numbers directly, contrasting 6% and 0.4% to argue nearly all attrition happens in the first week, guiding where to invest.
Proposes a concrete activation action (first meaningful interaction within session one) instead of vague 'improve onboarding'.

Proposing broad, unfocused fixes across the whole funnel instead of targeting the specific week where nearly all users are lost.
Setting an unrealistic goal like doubling week-3 retention directly without first fixing the much larger week-1 leak.

What would you test first if you had only one sprint to move the needle?
How would you know if the problem is onboarding versus a lack of relevant inventory for these users?