Product design question
How would you improve Twitter? What success metrics would you use to track your improvement?
- Product design
- Medium
Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.
Start a mock interview on this question · Mock interview from a job description
What this question tests
Tests prioritization and metrics thinking together, requiring the candidate to propose a specific improvement and pair it with a metric that actually measures success.
How to approach it
- Clarify what improve means: growing daily active usage, improving conversation health, or growing ad revenue, since Twitter's goals span all three.
- Identify a specific, known pain point: low-quality or toxic replies discourage participation and dilute the value of following accounts.
- Propose a concrete feature: better reply visibility controls and quality-based reply ranking, surfacing high-signal replies above low-effort or toxic ones.
- Weigh the trade-off: stricter reply ranking could suppress legitimate dissenting opinions if not carefully designed, a known sensitivity on this platform.
- Choose a success metric matched to the change: conversation participation rate (replies per active viewer) and a quality proxy like report rate on visible replies, not just raw reply volume.
- Note that raw engagement metrics alone (likes, retweets) wouldn't capture whether conversation quality actually improved.
What a strong answer includes
- Picks a specific, well-known pain point, reply quality and toxicity, rather than a generic more engagement answer.
- Names the real risk of over-moderation, suppressing legitimate opinions, showing awareness of the platform's core tension around free expression.
- Proposes a success metric specifically matched to the change, not a generic DAU number that wouldn't isolate this feature's effect.
- Uses illustrative numbers as assumptions: assume 20% of users report avoiding replying to popular posts due to toxic responses.
Common mistakes
- Proposing a feature with no clearly matched success metric, defaulting to vague engagement.
- Ignoring the free-expression sensitivity that makes content ranking changes especially scrutinized on this platform.
- Choosing raw engagement volume as the metric when the actual goal is quality, not volume.
Likely follow-up questions
- How would you define reply quality algorithmically?
- How do you avoid this being perceived as censorship?
- What would you do if the feature reduced overall reply volume?
More product design questions
- Redesign Twitter for eCommerce.Shopify · Product design · Hard
- Design a new feature for Twitter that improves the new user’s experience.Twitter · Product design · Medium
- How would you improve Twitter?PayPal · Product design · Easy
- How would you improve Instagram?Meta · Product design · Easy
- How would you improve the sign up process for LinkedIn?LinkedIn · Product design · Easy
- How would you improve Facebook Reactions?Meta · Product design · Easy
More questions from Twitter
Learn the skill behind it
Chapters of the AI PM course that teach what this question tests.
- Chapter 4: Discovery and strategy for AI products
- Chapter 7: AI UX and human oversight: design for a system that is wrong sometimes
- Chapter 14: Get the job: the AI PM interview loop