By Stuart Rose, global insurance marketing manager, SAS
There are no easy ways to grow revenue, increase profit or improve market share in the insurance industry right now. Increasingly, insurers need to look internally to trim expenses and improve business processes. One way to do this is to use predictive claims analytics to optimize loss reserves, increase productivity and assist in preventing fraud.Claims analytics is the process to analyze the structured and unstructured data at all stages in the claims cycle (first notice of loss to payout to subrogation) to make the right decision, at the right time to the right party. Rather than analyzing one case at a time -- based only on currently available information -- analytics gives you added perspective by allowing you to view this one claim "in context" by comparing it with previous claims settlements in your database.
The power of claims analytics is that it works in conjunction with your existing claims management systems. Whenever claims data is entered or updated, analytics can be used to reevaluate the claim for loss reserve amount, fraud or subrogation opportunities. The ultimate strength of those evaluations, of course, lies in the amount and quality of available data.
Claims fraud is already a widespread problem for insurers, and in a difficult economy it tends to accelerate. The most effective way to combat both opportunistic, like loss padding and organized claims fraud is to use a combination of business rules, exception reporting, anomaly detection, predictive modeling and social network analysis.
Business rules continue to be an effective means to combat opportunistic fraud. However, organized fraudsters actively test these rules and thresholds to find ways to operate around or under them. By using predictive analytics, insurers can compare a loss with similar claims and determine if there is significant deviation from the norm, usually an indicator of suspicious activity. Finally, by adding social network analysis, insurers can detect and prevent organized fraud by going beyond a single claim transaction view to analyze all related activities and relationships at a network level.
Another challenge insurers facing today is the inability to accurately forecast loss reserves and ultimately predict outcomes once a claim has been submitted. Using analytics it is possible to calculate an accurate loss reserve amount and benchmark each claim based on similar characteristics and hence reduce the propensity for loss padding. For example, data mining techniques have helped insurers identify that the size of a claim payout grows significantly based on the number of days between when the claim occurs and when it's reported. In most instances the size of a claim can increase by approximately 50 percent if the insured does not report the claim within the first four days.
Insurers often only receive a fraction of not-at fault settlement costs because they don't pursue subrogation opportunities. Among the problems are the sheer volume and type of data surrounding a claim. It is estimated that up to three-quarters of claims data is unstructured. Examples of unstructured data are emails, adjuster notes, medical records or police reports. Many recovery opportunities are missed simply because the indicator for a possible recovery is hidden in the claims narrative. Claims analytics help insurance companies find these opportunities by analyzing the textual information, identifying known subrogation characteristics and optimizing associated activities; therefore, loss adjustment expenses are lowered.
Finally some insurers are beginning to use analytics to calculate a litigation propensity score. Claims that involve an attorney often double the settlement amount and significantly increase an insurer's expenses. Analytics can help insurers determine which claims are likely to result in litigation, and mitigate those claims to more senior adjusters who can settle the claims sooner and for lower amounts.