Case Study: Coinmama reduces fraud review time with Pecan predictive analytics

Coinmama reduces manual transaction reviews by two-thirds with Pecan

Coinmama, a cryptocurrency exchange with over 3.5 million customers, faced the challenge of using static, manual methods to identify fraudulent transactions. Their custom SQL scripts were inefficient and limited to past variables, preventing them from detecting emerging patterns of risk. To evolve their analytics, they turned to the vendor Pecan and its predictive analytics platform for a solution in fraud detection.

Pecan implemented predictive models that automatically score thousands of transactions weekly for fraud likelihood. This solution reduced the number of transactions requiring manual review by two-thirds, saving an estimated 140 hours of analysts' time per month. The Pecan model also improved precision significantly and helped uncover 15% more fraudulent transactions, leading to better early detection and major time and cost savings for Coinmama.

Coinmama