You are a Data Scientist at a social app. The product team proposes a new in-app feature (e.g., a new sharing surface). You have event-level data and can run experiments. Answer the following: 1) **Using the data you already have (observational / pre-launch data), how would you identify whether users need this feature?** - Be explicit about what “need” means and what signals would support it. - Include how you would segment users and avoid misleading conclusions. 2) **Assume you can obtain any data you want. What additional data would you collect to answer (1) better?** - Include instrumentation you would add and any qualitative inputs you’d combine with quantitative data. 3) **How would you measure whether the feature is successful after launch?** - Define a North Star metric and 2–4 supporting/guardrail metrics. - Describe an A/B test design (randomization unit, duration, power/MDE considerations). 4) **How would you clarify the feature goal and trade-offs with stakeholders?** - Give examples of common trade-offs (e.g., engagement vs. quality, retention vs. spam/abuse, creator vs. consumer outcomes). 5) **How would you detect and handle novelty effects?** - Describe what you would look for in time series and what analysis or experimental design changes you would use to ensure the result is durable. Assumptions: users can generate multiple events per day; you can track impressions, clicks, downstream actions, and retention. State any additional assumptions you need.