Analytics Can Be Your Best Defense Against Corporate Fraud
Globally, organizations lose as much as $3.7 trillion to occupational fraud each year. The Association of Certified Fraud Examiners reports that theft, corruption, financial misstatements and other misdeeds cost the average business up to 5 percent of its annual revenues (bit.ly/1MIno0J). In the U.S. alone, on-the-job fraud, such as embezzlement, leads to an estimated $652 billion in losses annually, according to AllBusiness.com (bit.ly/1Nlo0t3). Adding insult to injury, 65 percent of organizations hit by fraud only recoup 25 percent or less of the stolen funds. They also suffer losses that are more difficult to quantify but no less impactful: lost productivity, reputational damage and even regulatory scrutiny.
What makes companies so vulnerable to fraud? With trillions of dollars at stake, organizations face a constant barrage of threats inside and out. Unheeded warning signals may open the door to fraudulent activity or, worse, allow wrongdoers to escape detection. Traditional approaches to prevention—whether external (e.g., firewalls) or internal (e.g., audits)—only slow down criminals. Even as companies develop more sophisticated fraud prevention tactics, criminals devise new ways to commit fraud and evade detection. Certainly, the losses to fraud would rise if organizations didn’t have basic protections in place, but rather than trailing behind, companies must evolve faster than those who perpetrate fraud.
Supplementing traditional fraud prevention methods with analytics improves an organization’s ability to pinpoint patterns or trends in data that indicate likely fraud schemes. By combining analytics, human insight and machine learning, companies can predict attacks, protect against improper transactions and take steps to prevent them in the future.
Predictive Modeling in Action
Because fraud is hard to detect and can happen quickly, organizations may not realize they’ve been defrauded until after the money is gone. Such was the case for Medicare when it paid nearly $35 million to a criminal cohort that used stolen personal information to bill the program for unperformed medical services, according to Analytics magazine (bit.ly/1WeaZmN).
Predictive modeling can play a significant role in the early warning, detection and monitoring of fraud, and early detection of fraud—before a payment is made—can significantly reduce losses. Recognizing its own vulnerability to billing swindles, the Centers for Medicare and Medicaid Services (CMS) implemented a system for identifying fraud using predictive analytics, which in 2014 helped the CMS identify or prevent more than $210 million in fraud, as noted in Modern Healthcare magazine (bit.ly/1SZvTky). By aggregating historical data of past fraud events and using those variables in predictive models, companies can anticipate the likelihood of future fraud events and take appropriate steps.
Similarly, connection analytics can identify individuals and companies suspected of engaging in fraud by offering greater visibility into fraud across a network, in addition to a single person or entity. By spotlighting such links, the CMS can identify new threats earlier, minimizing financial losses.
comments powered by