In recent years, growing losses from fraud have cost financial institutions millions upon millions of dollars. According to Visa International, gross fraud from debit and credit cards is up 29% in 2005 over 2004. TowerGroup predicts that fraud from debit cards alone will surpass $1 billion USD in 2006. Fraudsters have become more sophisticated, more organized and more international in scope than ever before. Financial institutions are under attack on a daily basis and continue to experience losses even with various fraud control mechanisms in place.

In response to this growth in transaction-based card fraud, most issuers have made use of neural networks and behavioral models to distinguish normal cardholder activity from fraudulent. While neural network technology and scoring models provide the base for nearly all fraud service programs, there are additional areas upon which issuers can focus to further improve their profitability by reducing fraud losses.

As effective fraud programs analyze data coming from multiple sources, be it the cardholder, the transaction, the merchant and/or card networks, it becomes increasingly important to create and deploy rules as a result of fraud pattern analysis into a real-time neural network. By using a real-time neural network engine, issuers are able to decline authorization requests for fraudulent transactions before approvals are provided back to the merchant. Near real-time systems receive transaction data after the transaction is authorized, which means that at the very best, only the second transaction can be declined. Since the vast majority of card fraud takes place within the first several hours of the fraud event, it is imperative that the processor analyzes data throughout the day, searching for fraud patterns or information that leads to adding to or modifying the rule set in place.

In addition, it is vital to conduct point-of-compromise analysis on a systematic basis to ensure that fraud trends are identified and stopped early. Some questions to ask: What type of point of compromise analysis is performed? How well is information leveraged across the portfolio? If your institution's card portfolio is processed by an outside party, how is the point of compromise information communicated to you as the issuer? Compromised card lists provided by networks contain useful information, but are often provided well after the event took place. Issuers clearly want to address both the immediate need of selectively reissuing cards as well as aiding in the investigative process, so the key is to identify and act on the information as quickly as possible.

Properly structured data warehouses and data mining tools aid tremendously in the effort to identify points of compromise early. Technology is not the entire solution, however. Skilled fraud analysts play an enormous role in identifying test transactions, merchants that may be colluding with criminals, and patterns where fraudsters are committing fraud once they have obtained valid cards.

The integration between the analytical side and the operational side of the solution is also significant. Fraud operations is an integral part of a fraud prevention solution, since that's the area where cardholder contact is made. Issuers can make a direct and immediate impact in reducing their fraud by ensuring they have current cardholder contact details. That has always been a struggle for many institutions, but will help to work fraud cases faster and minimize disruptions for cardholders conducting legitimate activity.

A fraud operations group also complements the analytical side of a professional fraud management solution by incorporating effective challenge and dispute procedures. This is an area where skilled customer service representatives are invaluable. It is not only important to verify fraudulent activity with the cardholder so that proper claims can be initiated, but it is also important to ensure that friendly fraud isn't included in claims submitted. This type of challenge is essential in keeping an issuer's net write-offs to a minimum.

Lastly, how effective is the investigative group? Is your organization or EFT processor well connected with local and federal law enforcement agencies? Does your institution work with other institutions to build a stronger case against fraudsters? The investigations and recovery are important not only for the immediate case your institution may be pursuing, but also because successful convictions send strong messages to others on the street. Criminals will always follow the money, but will start at places where they believe they will have the easiest time stealing.

Transaction-based fraud crosses issuer portfolios, PIN and signature networks and merchants across the globe. It is increasingly important that financial institutions employ expertise to manage fraud services built on 21st century technology. The ability to quickly synthesize data from multiple sources and take appropriate action will differentiate issuers who make use of effective fraud prevention programs. Organizations not engaged in fraud management, or who simply contract for a minimum set of fraud services in order to check the box, will begin to see their profitability dwindle.

Instant Insights /

Beyond the Neural Network: Analyzing, Investigating and Integrating Key to Fraud Management

Related Stories

Recommended For You