Case Study: Global Payments Processing Company achieves $30M annual savings and uncovers largest historical fraud with Cloudera

Global Payments Processor Finds Fraud with Hadoop while Saving $30 Million

A global payments processor was struggling with an expensive, resource-constrained enterprise data warehouse (IBM DB2) and mandatory disaster-recovery replication that drove up costs and limited analysts to roughly a year of data. Heavy ETL compute charges and the need for deeper fraud and risk analytics meant the security team could not run large-scale, long-duration queries needed to detect sophisticated fraud.

The company deployed Cloudera Enterprise Hadoop (about 300 TB on 52 nodes, ingesting 2→4 TB/day), offloading historical storage and aggregations to Hadoop and using Impala for low-cost, queryable DR. This enabled deeper exploratory analytics, uncovered the largest fraud incident in the firm’s history, cut ETL workload costs by 10–15%, sped up merchant reporting, and reduced DR/storage costs—saving about $30 million annually.