Tripling profit through fraud automation, acquisition economics and cheaper funding

Situation

At about a million applications a month, a lender serving underserved borrowers fights three battles at once: fraud, acquisition cost and the price of capital.

Approach

• Fraud: Moved fraud decisioning from largely manual review to 96%+ real-time automated decisions across roughly 1MM monthly applications. Built the fraud function from scratch.
• Credit: Optimized credit policy and deployed multiple machine-learning model updates.
• Acquisition: Built an affiliate lead-buying system that cut acquisition costs in half.
• Funding: More than doubled available debt capital while lowering funding costs.
• Built credit, fraud, finance and marketing functions with a team of more than 15 people.

Result

Company profits tripled over the period.

Relevance for clients

In high-volume, near-prime and subprime lending, the biggest gains often come from the edges of the credit decision rather than the model itself: fraud gates, lead pricing and the cost of funds. Each depends on the same underlying data, and each pays off faster when managed together.