A notable ecommerce company discovered that it could significantly increase fraud capture in real time by running AtData’s Fraud API as users attempt to create an account within their sales process. Not only did it reduce the volume of fake accounts created but it also used that data for future monitoring throughout the company.
This company experienced high fraud rates via their online sales which invariably led to high chargebacks, account takeovers, identity theft and a massive increase in fake accounts. Since the company’s fraud processes where overwhelmed, losses doubled as the fraudsters deemed this site an easy target.
In selecting their vendor, and talking with our fraud experts, they had to determine the right solution to meet their needs and scale. To determine which needs were the best fit, the company spoke with AtData’s fraud experts. Our team offered a comprehensive test solution so the client could quickly see the benefit and accuracy of the product prior to a full integration.
AtData was a perfect fit because the email was a key data point of the company’s account opening process and guest checkout. Therefore, in consulting with AtData this company could determine how and where to implement the solution to ensure maximum effectiveness.
How It Works
AtData’s Fraud Prevention service, at its core, is designed to provide a holistic view of the email and automatically assess risk.
The keys to that process are the following:
- Fraud Risk Score: Machine Learning models assess global meta data determine email risk
- Date first seen: The date the email address first became active
- Longevity: A score indicating when the email address was first encountered
- Velocity: A score describing the level of activity of the email address over the past 6 months
- Popularity: A score reflecting the popularity of the email across sources in the past 12 months
- Validation: Flags invalid and risky email addresses through proprietary methods
- Domain Check: Blocks toxic or known high-risk domains
- Tumbling Check: Catches fraudsters using multiple variations of the same email address
- Fraud Consortium: Identifies fraudulent emails reported by our network
Each of these metrics and data elements is critical in its own way for understanding if an email address is potentially fraudulent. Based on our feedback from clients, the date first seen of an email is typically a strong indicator of an email’s legitimacy.
That said, velocity and popularity have their place, too. When these data points combine with a client’s first party data/analysis, the client gains excellent insight into how widely used an email is, and, therefore, its legitimacy. This means fewer fake or malicious accounts, customer inquiries and purchases.
Most fraud prevention solutions are a custom fit based on the client requirements and its customers requirements. This ecommerce company emphasized its need to ‘say yes with confidence’ and quickly stopped the surge of fraud they were experiencing. The client fully implemented the Fraud API in its front-end registration page within 2 week and could immediately identify the bad actors creating fake accounts and subsequent credit card transactions.
Within weeks of using the AtData Fraud API, the customer reported dramatically lower fraud rates and attempted fraud rates because fraudsters could no longer succeed in creating accounts.
This in turn reduced the chargeback rate and the client was delighted to see significant increase in sales conversions and lowered friction for its not-at-risk customers’ experiences.