Case Study: Urban Outfitters achieves 20% reduction in order reviews with Accertify's machine learning

Urban Outfitters Cuts Fraud Order Reviews 20% with Machine Learning from Accertify

Urban Outfitters, a multi-brand specialty retailer selling via stores, catalogs and eCommerce, faced rising fraud as its online sales grew. The company relied on a large fraud-analyst team and high manual order-review rates to block fraud, and in 2015 set a goal to cut order reviews and false-positive cancellations while keeping fraud levels steady.

To meet that goal Urban Outfitters deployed Accertify’s Gradient Boosting Machine (GBM) machine-learning model for fraud decisioning. After implementation the company reduced order reviews about 20% across brands (peaking at a 51% drop during the holiday season), cut false positives by 8%, sped fulfillment, and significantly lowered fraud-review labor (shrinking the analyst team from 17 to 7), while improving risk prediction in international markets.

Urban Outfitters

Bryan Whitney

Contact Center Director