1.1. IDV Scores

ADARA’s identity verification framework is based on two key scores: Trust and Confidence.

1.1.1. Trust Score

The Trust Score evaluates anomalous behavior in a consumer’s digital activity in order to determine whether that person can be trusted for a given transaction. A consumer's transactional history across vendors paints a deterministic picture of that individual. Consumers follow consistent patterns of digital behavior, including visiting similar sites, from consistent IP addresses, and operating sessions in regular time intervals. The Trust Score identifies and monitors irregularities in those patterns for a given consumer.

The Trust Score considers a large number of features built to indicate anomalous digital behavior. The major components include:

• Amount and velocity of activity over time, in relation to other consumers. This is examined for common types of activities included page searches, purchases, and use of loyalty points.
• Amount and velocity of activity on associated digital identities over time. For example, a consumer who creates over a hundred new digital identifiers in one day would be demonstrating anomalous behavior.
• Origin of digital activity. The Trust Score evaluates whether a consumer uses multiple IP addresses and how often they switch identities and IPs.

These features and more are used to model the probability that a given consumer can be trusted.

1.1.2. Confidence Score

The Confidence Score measures a consumer through the lens of data availability, in terms of digital activity across activity types, verticals, and identifiers. This includes how broadly the consumer’s data can be found across the ADARA Data Consortium. High confidence scores indicate that a given consumer has demonstrated good online behaviors and made quality digital transactions across a number of different categories such as retail, travel, and finance.

In creating the Confidence Score, features are used that indicate both the health of an identity, and its proximity to or availability for digital transactions. For a given consumer, the score takes into account ID quality, availability, recency, frequency, and more. Features include:

• Amount and type of activity across verticals and categories within verticals. For example, within the travel vertical, a consumer who both purchases flight tickets and makes a hotel reservation will have a higher confidence score than a consumer who only searches for flights.
• Amount and quality of partner data where activity is shown.
• Number and types of identifiers associated with the consumer, indicating availability in the digital space.
• Recency and frequency of digital activity.

The resulting Confidence Score assesses how confident we can be in a given consumer.