How would you evaluate UberEats growth?

You have just joined UberEats as a senior data scientist. The interviewer asks you to reason about marketplace health and causal impact. Answer the following related product analytics questions: 1. **First-week diagnostics:** What are the most important metrics you would review in your first week to understand whether the UberEats marketplace is healthy? Your answer should cover demand, merchant supply, courier operations, reliability, and customer experience. Explain how you would segment the metrics and which guardrails you would monitor. 2. **Merchant overload blocking feature:** UberEats plans to add a feature that temporarily blocks new orders when a merchant already has too many orders in progress. How would you evaluate whether this feature is beneficial? Define the product hypothesis, experimental design, unit of randomization, primary success metrics, guardrail metrics, and key risks such as marketplace spillovers or interference. 3. **Incremental value of a new merchant:** Suppose UberEats signs a new merchant in a neighborhood. How would you estimate the *incremental* value created by adding that merchant, rather than just attributing all of that merchant's orders to growth? Discuss cannibalization, selection bias, and what causal inference strategy you would use if randomized rollout is and is not available. You may assume the marketplace has the following entities: - **Consumers** who browse, place orders, and may churn or retain. - **Merchants** with menu quality, prep time, backlog, operating hours, and cancellation behavior. - **Couriers** with availability, utilization, and wait time. - **Orders** with funnel stages such as impression, click, add-to-cart, checkout, accepted, delivered, canceled, and refunded. The interviewer is looking for a structured answer that balances short-term conversion with long-term marketplace health.