Predictive analytics: A growing trend in cost containment

Predictive analytics, or predictive modeling, is the ability to predict future probabilities and trends based on the measurement of a particular variable, and more and more companies are using this analysis to help them control costs. A good example of this is in the area of auto insurance, where insurance companies determine how much an individual will pay for their automotive policy based on their age, gender and driving record. With so much of total health care costs centering around specific portions of the population, predictive analysis is also valuable in identifying potential costs associated with a person enrolling into a health care plan. By looking at lab records, insurance claims and other information, a health care underwriter can determine how expensive it will be to cover that person based on these predictive variables.

The use of predictive modeling as a tool is a trend to watch in insurance both now and in the future. And there is plenty of evidence to suggest analytics are already well-established – and are here to stay.

In a 2013 survey of 269 insurance professionals commissioned by Earnix Inc. and Insurance Services Office, Inc., 81 percent of those polled said they routinely used predictive modeling for pricing, with 52 percent of those polled employing modeling for underwriting. Also, 50 percent of those surveyed undertook predictive modeling to investigate fraud.

Furthermore, the majority of insurance experts polled were committed to ensuring the data they collected was fit for predictive modeling, with 54 percent of those participating taking at least three months to collect and ready the data to be studied.

However, some questions remain in wondering if predictive modeling is the sliver bullet. As insurance industry expert Donna Popow once noted, tying a person's auto insurance rates to a credit score runs the risk of having someone pay a higher premium because of a non-driving related issue, such as a reduction in work hours.

Predictive modeling may not be for everyone, but most companies can benefit from it.  If you are wondering if predictive analytics can help your company or are looking to turn more possibilities into predictions through the use of these analytics, The Plexus Groupe, an innovative, client-focused insurance brokerage and risk consulting firm, can help. Visit Plexus on the Web at or contact us via telephone at 847-307-6100.


Cousins, Michael S., Shickle, Lisa M., Bander, John A. "An Introduction to Predictive Modeling for Disease Management Risk Stratification." Disease Management 2002; 5:157‐167.

Earnex Inc. and Insurance Services Office, Inc. "2013 Insurance Predictive Modeling Survey."

Popow, Donna. "Models Are Not Reality." Information Management, May 15, 2009.

Sclafane, Susanne. "Use of Predictive Models Widespread in P/C Insurance: Survey." Insurance Journal. November 7, 2013