Design metrics and an experiment for Eats donations
Company: PayPal
Role: Data Scientist
Category: Analytics & Experimentation
Difficulty: easy
Interview Round: Onsite
Uber Eats is considering a new feature: when a user places an order, they can optionally **add a donation to the merchant** (or a merchant-selected cause) during checkout.
You are the DS owning the evaluation.
1) Propose **primary success metrics** and **guardrail metrics**. Be explicit about definitions (numerators/denominators) and where in the funnel they are measured.
2) State 2–4 plausible **hypotheses** (including at least one potential negative effect).
3) Design an **experiment** to evaluate the feature (randomization unit, experiment population, variants, duration/power considerations).
4) List major **confounders/biases** or marketplace effects that could mislead conclusions and how you would address them.
Quick Answer: This question evaluates analytics and experimentation skills—specifically metric definition (numerators/denominators and funnel placement), hypothesis formulation, randomized experiment design, and identification of confounders—in the Analytics & Experimentation domain at a mid-to-senior data scientist abstraction level.