Ekata releases global identity verification dataset
|Brittany Hainzinger in Security Tuesday, June 16, 2020|
Ekata announced the release of the latest innovation to its global identity verification dataset, Network Score. Network Score is a machine learning prediction that enables businesses to better identify good and bad customers based on a series of insights.
Ekata announced the release of the latest innovation to its global identity verification dataset, Network Score. Network Score is a machine learning prediction that enables businesses to better identify good and bad customers based on a series of insights. The new dataset flags potentially risky digital transactions and fraudulent customers by analyzing the activity patterns of the identity information being used.
Network Score leverages the power of the Ekata Identity Network, a proprietary global dataset of billions of customer transactions, to reduce the number of false declines and increase the precision of fraud detection. The Identity Network works in conjunction with the Identity Graph, Ekata’s database of globally sourced and licensed data, vetted through rigorous acceptance criteria in compliance with global privacy and security standards.
“With over 20 years of sourcing identity data from our global data providers, we know that authoritative data isn’t enough,” said Rob Eleveld, Ekata CEO. “Stolen personally identifiable information (PII) and fake digital identities are becoming increasingly prevalent, which makes verifying identity in the digital and card not present (CNP) world harder than ever. Fraudsters can try to impersonate and act the way legitimate users do but they will never match 100 percent of the time; those activity patterns can be powerful signals of fraud.”
For instance, a real consumer typically uses the same primary address and secondary address together. But if we look at how that secondary address has been used across the digital interactions we’ve seen in our Network, we might see that the secondary address has been used with tens of different email addresses in the month, which suggests promotion abuse or other fraudulent activity. Ekata built the Identity Network to track these types of activities and leverage transaction-level intelligence to identify when consumer information is being misused.
The Identity Network, along with the Identity Graph, are unique, differentiated datasets that power the Ekata Identity Engine. The Identity Engine helps businesses make accurate risk decisions about their customers by providing predictive data insights on who they are, and how their information is being used online. Using sophisticated data science and machine learning, the Identity Engine transforms the two datasets into unique and valuable data attributes, such as Ekata’s new Network Score. These attributes are made available through Ekata’s APIs and SaaS-based tool, to vastly improve business’s confidence in their risk analysis.
Unlike other identity verification tools, the Identity Network offers dynamic decision making, as the model continues to learn with new transactions in order to better determine fraud potential. The Identity Network does not rely on blacklisting or a data consortium and does not use previous customer decisions to influence its data. Moreover, the Identity Network provides businesses insight into cross-border and cross-industry fraud patterns outside of their own data set.
Ekata has released a number of Network-derived features to help businesses maximize predictability in finding fraud, including:
- Network Score – Assesses the risk of a digital interaction using patterns and features selected based on predictive power in the Identity Network. Complex machine learning algorithms derive a score from activity patterns. Network Score boosts fraud detection efforts for real-time decision making and verifies genuine customers to help avoid false declines.
- Network Risk – Available through Ekata’s SaaS tool, Pro Insight, Network Risk equips manual review teams and fraud analysts with the top twelve signals that indicate positive or negative activity across a digital identity’s transaction history within the Identity Network.
- IP Risk – Uses a machine learning model to determine the risk that a given IP address is associated with fraudulent activity. IP Risk flags whether a given IP address is risky using statistical data within the Identity Network, as well as IP metadata such as Internet Service Provider (ISP), VPN proxy type, and more.