Some insurers see their data from new perspective

Five years can make a big difference in a company's perception of its data assets. In that time, Grange Insurance went from having its data in "kind of a mess," according to Tony Simpkins, the insurer's data warehouse project manager, to empowering its knowledge workers to solve business problems by seeing their data in a new way.

Like many in the insurance industry, Grange executives realized they were generating a huge quantity of data but were doing a poor job of leveraging it.

"We didn't have a data warehouse," says Simpkins. "We had several siloed databases we called a warehouse." Those databases didn't communicate with each other, and there was no easy way to generate reports across them, he adds.

Even when programmers cobbled together reports, the company's actuaries and business analysts complained that an item would be called something in one database and something else in another, even though they were describing the same thing. Or conversely, two things would have the same name in different databases but refer to entirely different things.

Another pain point was that anybody who needed reports written had to go through the IT department, which then came to be seen as a bottleneck.

To solve that problem, Grange began work in 2003 on a single, conformed data warehouse and turned to software company Microstrategy for business intelligence (BI) tools.

BI applications include data mining, online analytical processing and the creation of executive dashboards. Their goal is to help employees gain access to enterprisewide data and to provide tools to allow them to analyze that data.

Initially Grange execs thought the actuaries would use the BI software most, but the underwriting teams that deal with independent agents turned out to be the keenest users, Simpkins says. From their desktops, they can run 10 to 15 reports to share with agents and let them see where they're making their biggest profits. Agents who are having a problem can pinpoint where the problem is and come up with an action plan, whether it's re-underwriting some businesses or doing training. They can even drill down and see if the problem is with just one producer, and possibly recommend some training for that person.

"They could never do that before," Simpkins says, adding that the 250 Grange employees who are Microstrategy users are happy that they can get what they want when they want it.

Insurers play catch-up

As more insurers treat their data as a strategic asset, they are following Grange's path of cleaning up their data and using predictive analytics software to help them identify new business opportunities. But many have a huge task ahead of them before reaching the payoff business intelligence promises. Analysts and vendors say the sector lags behind others in the financial services field in its adoption of business intelligence. In terms of manipulating their data assets, insurers are seen as less nimble than the securities or banking sector, says Matthew Josefowicz, manager of the insurance group at research and consulting firm Celent.

Insurers are saddled with a complex data infrastructure, he explains. Systems may be divided by processes such as claims and billings, or by the range of products offered or even by state.

"So it's very difficult for them to get a consistent viewpoint across the enterprise," Josefowicz says. "There are companies taking important steps toward enterprise data models, but it's a reach for most in the industry to consolidate their source data."

Plenty of software vendors are offering tools geared to help insurers, Josefowicz says, but many insurers aren't in a position to take advantage of them.

"There's too much internal work that needs to be done before the vendor solution can add value," he notes. "The vendors who are most successful start from further back in the process on the data cleanup. If they start with ‘point us to your data,' that's often a problem."

Predictive analytics offers insurers richer insights into their data to price risk more accurately and to identify niche markets to target and others to push toward other insurers, says Chris Ciauri, senior VP of sales and marketing of Valen Technologies. His company's predictive analytics offerings target the property and casualty market. Valen helps customers improve pricing models and find previously unidentified market opportunities.

Ciauri gives an example of an insurance customer that wrote workers' compensation policies in Florida for many years. The general risk signature for Miami-area business was not very good and discouraged the company from writing policies there. But certain restaurant owners in Miami, who had been in business more than 10 years, looked like good risks at a certain price point, and the company's agents chose to aggressively target that group.

"So you may have an intuition that there's a pocket of business to pick up in an area where most people don't want to write," Ciauri says. "This gives you the statistical proof to back it up."

Typically customers start with underwriting, and when they feel comfortable with that they start looking at applications in other parts of their business, such as marketing and distribution, where predictive analytics might offer an advantage, Ciauri says.