← Back to Case Studies
Risk Scoring

Optimizing Fraud Risk Scores

North American Payment Processor ($5B annual volume)

🚨 Challenge

High false positives: 12% of legitimate transactions declined; yet >$1M/month fraud losses persisted.

💡 Solution

Built a composite risk score (doc, bio, device, phone, velocity); ROC‑curve thresholding; delivered analyst‑friendly explainability dashboards.

📈 Impact & Results

False Positives ↓
≈45%
Fraud Losses ↓
≈37% in 6 months
Analyst Productivity
≈2×

Visual Impact Analysis

Bar chart showing false positives reduced from ~12% to ~6.6%

False Positive Rate Reduction

Bar chart showing monthly fraud losses reduced from $1.0M to ~$0.63M

Monthly Fraud Loss Reduction

Hypothetical, representative engagement for illustration.

Talk to Us View All Case Studies