Design and evaluate P2P payments in messaging

Suppose you are deciding whether to launch a Venmo-style peer-to-peer payments capability inside a large messaging platform. 1) Give the single most important business objective (the objective function) and two to three concrete success metrics with precise definitions and units for the first 90 days; include at least one guardrail that protects core messaging health. 2) Suggest two monetization designs (for example, a per-transaction fee versus interchange/float) and explain how you would measure the causal effect on ARPU without letting subsidy-driven volume inflate the result. 3) Create a beta rollout plan that accounts for network interference and leakage: state the randomization unit (user, dyad, ego-network, or geo), define exposure, explain how you will handle access when a friend who was not assigned wants the feature, and describe how you will estimate ITT versus TOT. 4) If a randomized controlled trial is not possible, write the exact difference-in-differences specification (define treatment and control, pre and post windows, the estimand, the required parallel-trends checks, and how to cluster standard errors) and name two falsification tests. 5) Identify the top three risks—product, fraud/regulatory, and ecosystem—with a measurable leading indicator for each and a mitigation you would ship before full launch.

Overview: This question tests a data scientist's skills in product analytics, causal inference, experimental design with network interference, monetization measurement via ARPU impact, and risk mitigation for a peer-to-peer payments feature in a data scientist role. This prompt comes from a data scientist interview experience.