Provider & Fraud Knowledge – Square One
It’s tempting to approach fraud management in the same manner I assemble my son’s toys. Start at step 7, jump to 15, let out a few gasps, an occasional holler… and settle on whatever mix of parts you can jam together. Lucky for me, a toddler is often just as happy to play with the box as the actual toy. Businesses with online fraud problems aren’t quite as fortunate.
With respect to online fraud management, I find it beneficial to to revert back to square one. Learn about the different pieces and parts. Get a high level understanding of your options. And while there aren’t any one set of instructions, you will have a good foundation of knowledge and be able to move through your buying journey with confidence.
Here are some definitions for various fraud concepts and provider classifications. They hopefully lay the groundwork and/or fill in some gaps. As always, people vary on the specific language and groupings. This is ours. If you have additions/changes that you think would improve this list, we welcome all feedback.
Provider – Functionality
These solutions offer a fraud prevention platform that can manage fraud at various touch points, producing risk decisions and/or scores. These solutions are generally a hub for fraud prevention, solving a variety of use cases across the customer journey. Rules engines and/or machine learning are popular technologies leveraged by fraud platforms, deployed in a stand alone fashion or as complimentary to one another. While the core functionality is the same for this group of providers, these solutions vary greatly in technology, fraud/vertical coverage and other distinguishing features.
Identity & Authentication
These solutions focus on verifying digital identities and mitigating identity and account fraud. While this functionality is present in some platform & decisions engines, these solutions have a more targeted focus on authenticating users and protecting against fraud schemes like account takeover and social engineering. These approaches and practices often roll up into higher level methodologies like multi-factor authentication (MFA) and risk-based authentication (RBA) The latter leverages a variety of technology including biometrics, behavioral biometrics, device intelligence and machine learning.
Identity & Data Verification
These solutions also verify digital identities, but their core functionality is leveraging identity data at scale. Whether via API calls or web tools, these solutions verify numerous identity attributes to help businesses confirm legitimate customers at various touchpoints. They often focus on the legitimacy of identity data and whether it verifies to a known entity.
These solutions are dedicated to investigating and winning chargeback disputes through in-depth research and domain expertise. While some solutions manage chargebacks, these solutions differentiate as they provide a dedicated focus on this layer of fraud management.
Anti Money Laundering (AML)
These solutions analyze and monitor behaviors and transactions to prevent criminals from disguising illegally obtained funds as legitimate income. While leveraging similar technology as fraud prevention, AML solutions have a distinct difference in processes, workflows and regulatory mandates.
Technology & Data
The use of distinctive, measurable physiological characteristics to verify an individual’s identity. Physical biometric analysis includes techniques such as retinal scans, fingerprints and voice prints. Passive biometrics use behavioral data to authenticate, identifying anomalies to develop behavioral risk profiles. While subtle differences, passive biometrics are similar to behavioral analytics.
The analysis of how an individual interacts with a given device, whether that be a desktop browser, mobile browser or a mobile app. This includes interactions such as key strokes, scrolling patterns, tap pressure and many more. Profiling this behavior enables the solution to detect anomalies from genuine customer behavior and/or detect when behavior aligns with traditionally fraudulent behavior.
A subset of artificial intelligence, machine learning has the capacity to learn over time without being explicitly programmed. It can ingest a large amount of data and detect patterns and anomalies at scale. Machine learning models are trained to make risk decisions and/or generate a risk scores. For now, we are including supervised and unsupervised machine learning under the same category.
Fraud rules are algorithms that use specific attributes and parameters. A Rules engine allows for the creation and management of these fraud rules in order to make risk decisions and/or generate a risk score. While not self-learning in nature, rules engines enable analysts to test, modify and improve rule performance.
Identifying and tracking device activity by capturing and evaluating multiple device signals, commonly referred to as a fingerprint or DI print. Device registration and networks of coordinated device intelligence also assist in identifying and understanding device behavior.
IP & Geolocation
The utilization of an IP address, along with other device signals, to determine location and assess matches. Proxy detection rolls up into this broader category.
Consortium Level Data
The cultivation and utilization of large amounts of identity data to verify digital identities. This can be done through API calls, web tools or integrations into workflows and data models.
Fraud & Abuse Types
Unauthorized execution of any monetary transaction. This fraud spans multiple industries. In ecommerce it can include payments via credit and debit cards. In retail banking it can include money movement transactions like ACH, wires and P2P transactions.
Unauthorized access to a user’s account in order to steal identity credentials, execute a fraudulent transaction or engage in varying types of abuse.
New Account Fraud
Unauthorized opening of a new account leveraging compromised identity information. This can be for a variety of accounts, including credit cards, retail bank accounts, consumer lending and much more.
Synthetic Identity Fraud
A subset of new account fraud, synthetic identity fraud is when a fraudster combines real and fabricated identity data to establish a fabricated identity and open fraudulent new accounts.
Abuse of promotional offers by circumventing conditions/rules in order obtain significant discounts.
Abuse of loyalty points to obtain significant discounts or sell for profit. This abuse can include account takeover as a means of stealing loyalty points.
Abusive or malicious user-generated content. This abuse can include account takeover and/or new account fraud. Spam falls under this layer of content abuse.
Abuse of purchasing quantities in an effort to resell product for a profit. While reselling is a common practice, abuse can damage a client’s brand and/or deplete product availability for other customers.
Call Center Fraud
Exploiting call centers as a channel in which to launch fraud attacks, spanning varying forms of fraud and abuse. Account takeover is a common fraud type associated with call center fraud.