Metrics question

There are about 1 million inactive Netflix users. What would you do about them?

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 defining and acting on an inactive-user segment, distinguishing a real win-back strategy from simply flagging churn risk.

Clarify inactive: define precisely, subscribers who haven't watched anything in 30-plus days but are still paying, versus users who already canceled.
Segment the million: by tenure, device, and last-watched genre, since causes and win-back tactics differ by segment.
Diagnose likely causes: content fatigue, a forgotten subscription, a bad last experience, or a life change like a shared account no longer used.
Propose interventions: resume-where-you-left-off prompts, a curated recommendation email tied to their last-watched genre, and a pre-renewal reminder of what's next.
Weigh the trade-off: some inactive-but-paying users are pure profit if never re-engaged, so don't overinvest in win-back for likely churners regardless of effort.
Define success: reactivation rate within 30 days of an intervention, and reduction in cancellations from users who forgot they were paying.

Defines inactive precisely, paying but not watching, before proposing any action, avoiding a vague premise.
Segments the million users instead of applying one blanket win-back tactic to all of them.
Notes explicitly that some inactive-paying users are current profit, showing awareness that not all inactivity needs fixing.
Proposes personalized, content-specific interventions tied to each user's actual last-watched history.

Treating all 1 million inactive users as one homogeneous group needing the same campaign.
Ignoring that aggressively reminding forgetful-but-paying users could accelerate cancellations, hurting revenue.

How would you decide who not to target with win-back efforts?
How would you measure whether reactivation drives long-term retention or just a one-time watch?