Data Analysis to Identify Fraud, Other Suspicious Employee Behavior

A growing number of banks, retailers, and companies in other industries are making more extensive use of data analysis to help identify possible cases of internal fraud or other suspicious behavior by employees that can eat into profits and damage employer/employee trust.

For example, in the aftermath of the mortgage crisis, Rolling Stone made public emails that were written by Standard & Poor’s analysts as they inflated the ratings of worthless collateralized debt obligations.

“Let’s hope we are all wealthy and retired by the time this house of card(s) falters,” writes one S&P executive.

One approach that bankers and other executives can take is to use data analytics tools to monitor and flag transactions between customers and employees that are deemed suspicious.

For instance, many large institutions have implemented automated anti-money laundering (AML) transaction monitoring systems, according to a report by PwC. The monitoring systems are fed transaction data from various internal and external data sources.

These systems filter, compile, and summarize transaction data and flag instances of potentially suspicious behavior, typically by using rules-based scenarios to identify specific patterns of behavior by comparing transactional activity to algorithms that mimic known AML behavior patterns.

Another approach to flagging potentially suspicious behavior is through statistical profiling scenarios by modeling norms of client activity or segment activity over time and then identifying outliers as potentially suspicious.

The nature of employee theft is also changing, forcing retail, bank, and other industry executives to keep their guards up, according to a report by Aberdeen Group. Fortunately, predictive analytics tools can help retail executives gain a much deeper understanding of the sources of fraud and theft (store-level, warehouses, distribution centers).

Some of the tools that retailers are now using to identify and pinpoint suspicious activities include exception-based reporting systems that examine cash shortfalls and overages; suspicious returns; fake employee ID numbers; and other aspects of point-of-sale and retail activities, according to Datanami.

Predictive analytics solutions that incorporate real-time data streams enable retail, bank, and other industry executives to spot fraud as it’s occurring and respond quickly.

Next Steps: