Current Solutions Are Ineffective.

Neural networks and Bayesian inference are at the heart of many of today’s older risk management engines. Training these models takes time and huge amounts of transaction data. This means they lack the ability to respond to emerging real-time threats and usually only apply to credit card payments.

So-called “machine fingerprinting” technologies attempt to track the use of specific devices used by criminals. However, dual-boot machines with multiple browsers make device uniqueness a thing of the past. The widespread use of open proxies and compromised computers connected to botnets make it nearly impossible to track a crook by device. And if the user has JavaScript turned off, the machine parameters cannot be collected.

Kount Works—And Here’s Why:

Optimized specifically to detect suspect web transactions, Kount fraud detection incorporates real-time global assessments of transaction attempts across multiple:

• Merchants
• Customer Identities
• Machines
• IP Addresses
• Networks
• Geographies
• Methods of Payment

Kount risk assessments are far more accurate than older card-centric technologies that rely on typically stale data models and card-present usage patterns. And it works with all payment types including emerging brands like PayPal and Bill-Me-Later.

Kount Features

• Sub-second real-time reply with available email address correction service.
• Kount Dynamic Scoring^TM gives ongoing risk assessments until fulfillment time.
• Proxy Piercer^TM identifies customer network type (prison, library, high school, anonymizer) and geolocates customer behind the proxy.
• Kount DMC^TM is the Decision Management Console that gives merchants control over key business rules.
• Robust manual review automation tools to prioritize cases and improve efficiency.