Case Study: Traveloka achieves 2x more accepted orders and 3x less 3D Secure traffic with Sift

How Traveloka increased real-time bookings and stopped ATO attempts

Jakarta-based Traveloka, Indonesia’s leading travel booking platform, faced rising payment fraud from stolen cards and account takeover (ATO) attempts as it scaled across Southeast Asia. Its rules-based fraud system generated many false positives that blocked legitimate customers, harmed conversion and trust, and couldn’t keep pace with evolving fraud patterns.

Traveloka integrated Sift’s machine-learning platform, deploying two custom behavior-based ML models (one for payments, one for ATO) and using the Sift Console for investigations. Within weeks detection improved: accepted orders doubled, traffic to 3D Secure fell 3x, ATO rates remained low despite growth, and the team became faster and more effective at handling incidents.

Traveloka

Wayan Tresna Perdana

Sr. Product Manager