Half and Half

Getting 50 percent of a target group to achieve good-to-excellent results from a software tool might be the goal for some product providers. However, the value of business intelligence in the insurance industry is such that 50 percent seems like a low percentage to an industry analyst such as Matt Josefowicz.

The percentage of carriers doing good-to-excellent work in BI, he believes, should be higher than the 50 percent that responded in that manner in the study completed by Novarica, where Josefowicz serves as director of the insurance practice.

“The fact that only half said their business intelligence capabilities were good or excellent means half the industry doesn't have good BI capabilities, and that's a big number, especially when the insurance industry is so dependent on analyzing information to make decisions–that's at the very core of the business,” says Josefowicz. For many insurers, “there is a lot of discipline around business intelligence from the underwriting side. There's a lot more information available for underwriting and things such as marketing and optimizing operations. But still there are a lot of [different] decisions that need to be made.”

For other insurers, underwriting and new business are where the use of business intelligence tools is lagging, indicates Josefowicz. One reason for this is BI strategy is driven by a reporting orientation. “Business intelligence tends to start with financials, and it takes a while for the discipline to move forward to underwriting and new business,” he says. “Underwriting has its own sense of itself as having good data-analytics capabilities because that's what underwriters do–analyze data–but the actual capabilities lag behind what is possible now.”

A good sign from the study, notes Josefowicz, is another third of the insurance community is expecting more impact from BI over the next 12 months. However, some patience has to be shown, he adds, because in a lot of cases these are recent investments and new capabilities being developed.

One of the major impediments to a successful business intelligence program is an unwillingness of business users to change. “That's a serious change management issue and a leadership issue to make insurance companies information based in terms of their decision-making,” says Josefowicz.

It always is difficult to get business users to do their job in a different manner than was done in the past, Josefowicz continues. “You have built a culture around a certain assumption of what is a good type of customer or a bad type, what is a good risk or a bad risk,” he says. “It's hard to change [those feelings] just because a report comes along and tells you all your assumptions are wrong. You have a bunch of ingrained behavior based on those assumptions, and business structures based on those assumptions, so it takes a certain amount of change [management] to change that process.”

Compounding matters, data and analytics quality for some insurers have been so poor that in the past many executives didn't trust the information they were given, asserts Josefowicz. “[Executives] learned to go with their gut or go with other decision-making methodology because of not having a lot of faith in the data being presented to them,” he says.

Getting the most from a business intelligence solution is more than a problem for the IT department, points out Josefowicz. “It's not just a matter of data quality and building reports,” he says. “It's a question of what companies are going to do differently with the information. That's something that has to go into the business cases for the BI projects.” Business intelligence by itself is not going to increase sales volume or reduce underwriting loss, concludes Josefowicz. What BI does is give business executives the information they need to make better decisions. “Part of the business case for any of these advances has to be what business executives might do differently if they had better information and how they would act differently,” he says. “[BI solutions] can get you better information, but don't just say you need better information without tying [the request] to some kind of output.”