Choose target customers and define success metrics

Stripe plans to launch “Instant Payouts+” for SMBs. How would you choose initial target customers and measure success?

• Targeting: Propose a scoring model that ranks merchants using recency/frequency/GMV, payout latency sensitivity (e.g., weekend volume share), risk tier, support tickets about payout timing, and integration readiness. Define a threshold that yields ~10% coverage while maximizing expected incremental profit.
• North-star and guardrails: Define a primary metric (e.g., net incremental gross profit per merchant) and guardrails (chargeback_rate, dispute_loss_rate, churn). Specify exact 28-day measurement windows and baselining.
• Experiment design: Recommend a geo/merchant-level randomized rollout with stratification by country and industry. Given baseline adoption 5% and MDE +1 percentage point (to 6%), α=0.05, power=0.80, compute the minimum sample size per arm and discuss variance reduction (CUPED) using pre-period adoption.
• Leakage/bias: Identify two confounders in targeting (e.g., sales outreach) and how you’ll prevent them (e.g., holdout within high-score deciles).
• Operationalization: Instrumentation you’d add, success thresholds to exit the experiment, and a staged rollout plan if guardrails breach. Provide concrete formulas for lift, ROI, and sample size; include any assumptions you make.