Big Data = Big Challenge

A great deal of attention is being paid to "big data," as much of the unstructured data insurers compile is being called these days. Such data, amassed through tools such as telematics and social media, threatens to swamp companies thatdon't have the infrastructure in place to handle it and the analytics to make sense of it all.

But one problem many insurers face is they are still trying to put some order in the data they already have compiled from traditional sources or purchased from third-party providers.

"Certainly the right order of things is to get a handle on little data first, which means get a handle on your enterprise data and the kind of structured data sources you can use to improve underwriting and claims," says Matt Josefowicz, partner and managing director of Novarica and the author of the new report: Analytics and Big Data at Insurers.

In preparing the report, Josefowicz reports he noticed that even though there is a split between the analytics capabilities seen in large firms and small firms, those capabilities currently are not as pronounced as some might expect them to be.

"But when you look at future plans, the large firms are uniformly more aggressive in terms of planning to build out analytics capabilities over the next 18 months," he says. "There is much more variance in whether the smaller firms are going to be active in that area or not. I think we're going to see a real analytics gap between the companies that have made these investments and are in a position to use data more effectively and the companies that haven't [made the investments] and are going to use more traditional methods. Those insurers [in the second group] are going to be under more severe pressure."

A carrier's ability to deal with its own data is not necessarily a good indicator of any success that same insurer might experience in dealing with unstructured data.

"It depends on what type of big data they are looking at," says Josefowicz. "In the scaled and the unstructured data sources, there are different technologies required. There are newer data analytics and data management technologies that are designed to meet the challenges of the traditional, relational databases and traditional analytics rules thatdon't perform at the scale of big data."

Josefowicz believes the good news for mid-tier and smaller carriers is a class of third-party solution providers will evolve that will perform the analytics externally and provide the model and output to smaller insurers that won't have the capability to develop the infrastructure themselves. However, that segment of the market might not be able to match the timeframe that some large carriers have already established.

"More large insurers are building out the technology infrastructure for big data over the next year or so," says Josefowicz.

For certain large firms, preparing for big data needs to be a priority today, explains Josefowicz. For small firms, he maintains it will become a priority when they see they are being adversely selected again.

"If the larger firms are able to leverage new data sources and new capabilities to price more effectively and to segment the market more finely—cherry-pick the market through better underwriting, marketing and targeting—smaller firms are at risk of seeing the quality of their customers erode," he says.

One of the most discussed areas of big data has been telematics. Josefowicz believes there is great potential for the technology, but maintains the question to be answered is how can insurers take their telematics data and turn it into a pricing differential or an experience differential that's meaningful for the consumer.

"What's gotten lost in the telematics discussion is the pressure from the direct writers and the competitive environment among the direct writers that has pushed the price of personal auto policies way, way down," he says. "If I'm a personal auto consumer, how much more of a discount am I going to get by taking advantage of one of the telematics programs because the pricing is already so competitive in the marketplace?"

Josefowicz believes telematics will improve underwriting capabilities on the fleet side, but on the personal auto side the additional complexity that big data is emblematic of needs to be managed behind the scenes and not pushed out to the customer.

"The trick for insurers will be to use big data and other new analytics capabilities to simplify the customer experience, not to make it more complex," he says.