Catching fraud at the branch level

Despite headlines about cybercriminals stealing money across continents, statistics indicate that old-school, in-branch (employee or visitor) fraud remains a problem for many banks.

In fact, according to the Association of Certified Fraud Examiners' 2014 Report to the Nations on Occupational Fraud And Abuse, the U.S. accounts for more than 48% of all occupational (employee) fraud reported by victim organizations, and the banking and financial services industry has the highest percentage (17.8%) of fraud cases across all industries. Banking and financial services companies also have the tenth-highest median loss, at $200,000 per victim organization (the U.S. median is $100,000).

The situation with check and account fraud on the account-holder side is no less risky, due at least in part due to the obsolescence of many branch processes. Celent supports this viewpoint, with Banking Group Senior Analyst Bob Meara stating that banks not offering remote deposit capture (RDC) solutions are more susceptible to returned-item loss due to their outdated branch deposit-taking workflows.

Fortunately, much of the information that banks need to combat fraud, check scams, and other related activities is already in the branch, often in the form of unstructured or unprocessed data. Companies in all industries hear about the challenge of Big Data and the difficulties harnessing it, but few have access to the sheer volume of information that banks can leverage.

One of the hottest trends of 2014 is using business intelligence to combat fraud with proactive data analysis, interpretation, and reporting. Banks are perfectly poised to launch these initiatives in their branches, as well.

Data analysis and fraud

Most banks or their processing partners already use some form of data analysis to combat fraud, especially to identify fraudulent ATM/debit or credit card transactions. Now, analytics-based detection efforts are expanding to other types of transactions, such as wire transfers.

In March 2014, for example, IBM announced Counter Fraud Management Software, a new product designed to use Big Data analytics to examine multiple streams of data and help identify potential fraud perpetrators. This product, which targets the financial, healthcare, and insurance industries, is designed to analyze data related to a variety of activities where fraud is prevalent, including wire transactions.

Solutions like this one represent another dip into the Big Data pool, but most of the analysis is taking place at a high level. To date, in my view, few banks have placed sufficient emphasis on combating fraud by applying data analytics to activities inside the branch. Various experts have suggested that this may soon begin changing, and based upon the fraud statistics, we believe it cannot happen soon enough.

Where branch data and fraud intersect

Branch-level data collection and analysis can be effected nearly anywhere that a branch interacts with its customers or its personnel.

Lobby sign-in

The lobby is the first place this intersection occurs (and banks have the opportunity to collect data), yet many banks persist in using paper-based lobby sign-in systems. For fraud detection, lobby sign-in information provides a record of anyone who has visited each branch.

Banks that abandon paper sign-in sheets and embrace business intelligence solutions to collect and analyze this data on a bank-wide basis can better identify individuals that may have tried to perpetrate frauds at other locations.

Customer assistance sessions

Customer service encounters with account holders or prospects are often missed opportunities, as well. Without targeted data collection in place, customer service representatives may create or pull up account records, but they may not be able to record sufficient detail.

Whether branch representatives are helping a customer open a new account or gathering details regarding their need for a loan, data harvesting can be significant with proper technology, and all of that information feeds into the total data pool.

Financial activities

In addition, a variety of technologies exist that can capture data (tagged by processor and timestamp) from the financial activities that transpire in the branch, including transaction processing and teller balancing. This information is especially valuable for pinpointing occupational fraud such as skimming (short time frames between deposit and withdrawal are often evidence of this type of fraud).

Out of the silo and into the solution

The value of these data streams often lies, not in their individual databases, but in how banks can use them collectively to detect anomalies, develop behavior profiles (and then identify when activities do not match), and perform other analyses. In doing so, banks can glean a broad array of targeted information, provided through reports, alerts, and other mechanisms, that identifies the red flags of fraud.

Furthermore, the solutions that provide this analysis don't have to be SIEM (security information and event management) platforms like those used by major banks to obtain alerts about specific triggers within the system.

SIEM platforms can play an important role in fraud detection, but they usually don't natively integrate with business intelligence platforms―and BI-to-SEIM integration is both complex and expensive.

Innovative institutions are having success with smaller initiatives―either retargeting the data from existing solutions they have deployed for other purposes or implementing less-expensive, more narrowly focused tools and solutions.

For example, my firm recently worked with a financial institution which had lobby tracking solution that enabled it to identify attempts to open bogus accounts. Although the system was designed to help improve customer service and cross-sell ratios, the company was able to use sign-in and customer service representative session data―which the solution aggregated and analyzed for all branches―to run an additional check of individuals trying to open bank accounts.

To achieve this goal, customer service representatives were instructed to check each new account requestor against the database to see if they had attempted to perform questionable activities at other branches. The system thwarted some fraudulent attempts to open bogus accounts, and the institution continues to check all new account requests against historical records in the same way.

Moving to a new perspective

Across the board, experts point out that launching any BI initiative―even one that can solve crimes―requires more than technology. It requires a perspective shift on the part of management and personnel who, after all, drive (and provide the input for) the process. In a recent blog, "The Big Data Analytics Mindset," Gartner Research Director Anton Chuvakin expressed this point well, noting, "The only path is to shift the thinking, learn to be data-centric and data-driven and then solve problems that call for bigger data."

Furthermore, this mind shift happens best when positive results come quickly―another argument for starting small and eating the business intelligence "elephant" in little bites. The authors of the FICO Banking Analytics blog recently predicted that this year marks a watershed for embracing this approach, writing, "2014 will be the year in which, in the banking sector, innovation will move away from Big Data infrastructure towards analytic tools and applications." 

About the author

W. Michael Scott is president and CEO of Financial Management Solutions, Inc. The company provides business intelligence and performance management systems for branch-level staff scheduling and lobby management.