Context
A new digital lending product for a bank’s mobile and web channels, connecting customer experience to income visibility, credit policy, pricing and unit economics.
Problem
A higher limit is only safe if the system can establish affordability and apply the rules the same way every time. Product design had to work around the evidence available.
System and insight
Customer and application dataIncome estimationCredit bureau and policyDecision rulesProduct configurationMobile offer and application
Affordability is not a field on a form. It is a decision built from evidence, assumptions, policy and a consistent implementation.
Trade-offs
- Alternative data widened who could be assessed, but each new signal had to earn trust against the evidence already in use.
- Treating Zeus as its own product configuration gave freedom on pricing and limits, at the cost of one more product to run.
Intervention
- Redesigned the system architecture and brought in alternative data.
- Specified income estimation, credit logic, a concept card, business requirements, pricing and unit economics.
- Designed API-based microservices for offers, the decision engine and repayments.
My role
I led the project: architecture, requirements, product logic and delivery. Engineering, data science and risk teams built and ran it.
Result
- Project Zeus delivered 60% month-on-month revenue growth in its first year.
Learning
Ship the smallest system that can learn, then widen it.