The explosion of social media has opened a vast new trove of information — true and false — for people to express themselves and learn about the world.
In doing so, it has created an entirely new channel of communication for scammers that perpetrate fraud, and also for insurance companies to discover fraudulent claims.
For instance, consider the case of a California corrections officer who filed for disability benefits due to a back injury. Two days later, he participated in a mountain bike race, a fact that was discovered after Facebook posts, his headcam videos and race standings appeared online. He served 45 days in jail and paid a $5,000 fine.
In another case, a newlywed collected $26,500 after claiming she lost her wedding ring while swimming in the ocean. Suspicion arose when, a few months later, her husband also reported losing his ring valued at $14,000. Soon thereafter, an investigator found the wife’s picture on Facebook showing her with the ring. The insurer recovered its money.
The eternal battle between fraudsters and their victims plays out daily in the industry, since some studies suggest that as much as 10 percent of all property and casualty claims contain fraudulent elements, according to the Insurance Information Institute. Social media investigatory techniques must continuously improve to stay a step ahead of those with malicious intent, since perpetrators constantly develop new angles of attack.
The sheer volume of information now available, however, poses both the greatest opportunity and challenge for insurance companies. One billion new social media posts appear every two days, according to consulting firm Infosys’ report “Big data analytics: New whistleblower on insurance fraud.” How can a company possibly digest all of this information? The answer involves teamwork, a clear approach and big data analytical tools.
Here are eight tools and techniques that should be in every insurance company’s arsenal:
The right team
Leveraging social media to fight fraud requires a unique breed of investigator who can understand how the data available fits into the context of a larger investigation. They must be proficient in adhering to legal and procedural requirements, identifying subjects correctly and examining case documentation and public records before starting a social media search.
Due to the highly specialized nature of this role, forward-thinking organizations often engage select vendors that maintain teams of investigators with social media search capabilities.
To avoid getting lost in research blind alleys, social media investigators must choose what questions they want answered and be mindful of where they may be able to develop additional leads. For example, did an accident happen as reported? Is a business still operating successfully? Who or what is the subject connected to and what may that connection reveal?
This tool can parse and analyze unstructured text-based information. It has already been helping investigators examine internal claims data such as adjuster notes, emails, service calls and interview records. Top claims operations are using text-mining tools to search Facebook, LinkedIn, YouTube, Twitter and other social media sites.
Catching fraud in its earliest stages is one of the best ways to minimize loss, especially if it is discovered before significant funds have been paid. Predictive modeling can act as a powerful aid in supporting this objective. It looks for common patterns that indicate an increased possibility of fraud, allowing insurance companies to investigate sooner and often achieve a faster case resolution rate.
Social network analysis
SNA sifts through millions of social media posts and platforms to uncover hidden relationships among people, places, locations, accounts or virtually any other type of entity. For example, SNA can reveal when multiple claims all originate from a single household, or come from associates of a known crime ring. Perhaps the search reveals that a fire insurance claimant had been posting inquiries about selling the property immediately before the fire, or that a specific vehicle had been involved in a number of otherwise unrelated claim incidents.
To aid this analysis, investigator teams utilize link analysis software that illustrates connections and the development of relationships over time. This software capability not only reveals the whole picture but also cuts down the time needed to organize the information and identify the common factors between claims.
Related: The changing face of fraud
Boolean search terms and search sequencing
Advanced use of Boolean search terms can greatly improve the precision of searches and help account for name and other variations that can surround a person, object or circumstance. In addition, consideration of the sequence of different searches when tracking the subject or connections through the internet can factor in the quality of data discovered.
Cross-leveraging data sources
Recent approaches place considerable emphasis on interconnectivity between online information platforms such as social media, traditional websites, public record databases and vehicle telematics. For instance, fraudsters claiming they were rear-ended may be confronted by telemetric data showing the victim’s car was standing still at the time of impact.
The prospective data source evaluation process typically begins with an in-depth assessment of the person or business in question, as well as relationships between their connections to other people, businesses, groups, vehicles, properties, etc., and the likely platforms that will hold pertinent information and leads.
Near real-time monitoring
Powerful new tools allow the very best service providers to actively monitor for changes in the subject’s profile or internet activity, while assuring a high confidence level in the results by removing false positive findings. This means that conducting a social media search may not be a one-time task but something that needs to be done periodically or as events unfold during the life of a claim.
As a fraud-fighting tool, social media is a welcome addition to the many resources available to investigate claims. The threat of fraud never stops, and its perpetrators constantly evolve their tactics to find new angles of attack. Finding new methods to fight fraud is critical to keeping premiums down and policies affordable.
Brian Wilson is vice president, special investigations with QBE North America, where he is responsible for strategy development, resource and business model management, claim fraud investigation oversight and compliance support. He is a member of QBE's global counter-fraud team and an advisor to senior management.