How would you improve YouTube's recommendation algorithm
without increasing watch time at the expense of user wellbeing?
About this question
Category
Product Design
Subcategory
Improvement
Difficulty
Medium
Est. time
40 min
What this question helps you practice
This prompt tests whether you can balance recommendation performance with long-term user value. A strong response should define the wellbeing trade-off, segment user intents, propose responsible ranking improvements, and choose metrics that do not blindly reward more watch time.
How to practice
Define the objective carefully
Clarify what better recommendations mean beyond time spent: relevance, satisfaction, learning, trust, or control.
Segment user intents
Separate entertainment, learning, news, music, kids, casual browsing, and intentional search contexts.
Propose ranking and UX safeguards
Consider feedback controls, diversity, session boundaries, quality signals, and long-term satisfaction metrics.
Strong answer signals
Explicitly names the risk of optimizing short-term watch time at the expense of wellbeing.
Uses counter-metrics such as regret rate, not-interested actions, survey satisfaction, and harmful binge patterns.
Balances user value, creator ecosystem, business outcomes, and platform responsibility.
Common mistakes
Treating watch time as the only success metric.
Suggesting generic AI improvements without explaining ranking signals or guardrails.
Ignoring creator incentives and ecosystem-level effects.
Ready to try it?
Practice your answer in the interactive workspace.
Keep practicing
Explore frameworks and related prompts that build the same interview muscle.
