Business Context

You’re the analytics lead for Stripe Treasury’s B2B payments and settlement product. The platform processes $25B/month in cross-border payouts for marketplaces, payroll providers, and SaaS platforms. To expand into higher-volume corridors, the business wants to onboard new payout counterparties (e.g., regional banks, liquidity providers, and payment processors). Each counterparty introduces credit and settlement risk: if they fail to settle, Stripe may need to cover obligations to customers, creating direct losses and regulatory scrutiny.

Over the last two quarters, the Risk team observed that while onboarding volume increased 30%, losses from counterparty settlement failures rose from $1.2M to $2.1M per quarter. The Head of Risk believes the current onboarding checklist is inconsistent across regions and wants a single, metrics-driven creditworthiness framework that can be monitored weekly and used to set: (1) approval/decline decisions, (2) initial exposure limits, and (3) ongoing monitoring triggers.

Metric Scenario

You are asked to propose a primary KPI that summarizes the creditworthiness of a new counterparty at onboarding time, and a set of supporting leading indicators that can be monitored post-onboarding. Stakeholders include:

• Risk: wants fewer defaults and better early warning signals.
• Sales/Partnerships: wants faster approvals and fewer false declines.
• Finance: wants predictable loss rates and capital allocation.
• Compliance: needs explainability and auditability.

You have historical data for ~1,800 counterparties onboarded over 4 years, including outcomes (defaults, late settlements, limit reductions) and financial/operational attributes at onboarding.

Requirements

1. Define a single “creditworthiness” KPI suitable for onboarding decisions (not a model score in the abstract—define it as a measurable business metric).
2. Provide a calculation approach and explain how you would handle:
   • Missing financial statements (common for smaller processors)
   • New geographies with limited history
   • Counterparties with rapidly changing volumes (seasonality)
3. Decompose the KPI into actionable drivers that map to real interventions (e.g., tighter limits, collateral, shorter settlement windows, enhanced monitoring).
4. Propose benchmarks/targets for what “good” looks like and how you would set them using internal history and external industry references.
5. Recommend guardrails to ensure the KPI doesn’t optimize risk at the expense of growth (e.g., approval time, partner conversion).

Data Available

You have one week to deliver a written proposal and a dashboard spec for the quarterly risk committee.