Recently, James Taylor, president of Decision Management Solutions, in cooperation with SourceMedia, published the results of a
The insurance industry holds a special place in its heart for predictive analytics. Predictive analytics is all about understanding risk, and enables enterprises to either be alerted to trouble down the road, or to sidestep it altogether. Conversely, predictive analytics means identifying opportunities before they become apparent to anyone else.
Recently,
It's worth it, then, to take a deeper dive and drill down into some of the findings.
For example, the use of predictive analytics in pricing and ratings is fairly commonplace, with 71 percent of insurance companies either always or frequently using the tools for this purpose. But predictive analytics also is increasingly being employed in underwriting—used here on a constant basis by 52 percent of insurers. Here, the technology is primarily used in risk selection.
Personal auto and homeowners will be the focus of new predictive modeling initiatives the in next two to three years (49 percent each). Commercial auto (37 percent), commercial property (37 percent) and business owners (24 percent) also were mentioned.
One-fourth of insurers also have predictive analytics tools at work in their marketing departments, suggesting that there is quite a bit of cross-selling, target marketing and testing results across the various interactive and media channels. About 21 percent of claims departments also are using the technology—likely as a tool for fraud detection.
What kind of data goes into insurers' predictive analytics? Nine out of 10 insurers say they pull data from outside sources to supplement their own data. Of the companies that use external data, the vast majority (80 percent) use insurance score or raw credit attributes. Sixty-seven percent of respondents use geo-demographical data and 53 percent use competitive pricing data. Forty-six percent and 42 percent of the respondents also mentioned catastrophe model data and weather data, respectively.
While data quality is the issue that gets in the way the most of predictive analytics, insurance executives also are just as likely to cite a lack of skill sets to help build and refine predictive models. Close to half say they can't find enough people with the skills needed to assist in these efforts. This is where the challenge will be felt most acutely over the coming years. Insurance companies are competing with banks, health care establishments, manufacturers, government agencies, tech companies, web companies and everybody else for professionals who understand how to build and draw insights from predictive models. The survey finds it may take up to six months to extract and prepare data for modeling, so it's not a job that can be slam-dunked over a few days.
The cost of the technology is coming down, but the cost to keep good people in the business will be on the rise. This will increasingly shape the challenge for insurers that seek to compete on analytics.
Joe McKendrick is an author, consultant, blogger and frequent INN contributor specializing in information technology.
Readers are encouraged to respond to Joe using the “Add Your Comments” box below. He can also be reached at
This blog was exclusively written for Insurance Networking News. It may not be reposted or reused without permission from Insurance Networking News.
The opinions of bloggers on
