Although the insurance industry is quickly adapting to become more sophisticated in its use of big data, many companies still consider themselves to be behind the competition, panelists say at Standard & Poor's Rating Services' 30th Annual Insurance Conference.
For an industry that relies on information and actuarial assessment, big data has brought forth substantial changes in "variety, velocity, and volume," according to Murli Buluswar, chief science officer at AIG.
"While insurance has been a fairly data-rich industry historically, at the crux of it, we're trying to better understand our customers and bring them products that meet their needs," Buluswar said, adding that big data enhances users' risk mitigation, not just risk management.
But according to conference attendees, some insurance companies are lagging behind. According to a poll at the Standard and Poor conference, 42% of respondents assessed their firm's effectiveness in using big data as "lagging but hoping to be more competitive," while only 7% claimed that their companies were "leading the field."
According to Stan Smith, head of analytics at Milliman, a consulting group that helps companies better use data, there are two keys to effective data use. Ensuring that the right information is being collected, and the weaving together of various data points is done properly to separate "signals" from "noise."
For some insurers, the challenge in implementing more sophisticated big data strategies stems from ambivalence in breaking company habits that may have worked well in the past, but are now less effective. But according to John Anderson, head of strategy and transformation for Global Business Services at IBM, the potential for a company to transform its practices is clear.
"What used to take three months, you can do in 24 hours," Anderson said. "You're beginning to see associations between things that you would have never known about. The idea is to put the information in the hands of the underwriters and they use it to make big decisions. Theoretically, you could take big data and drive toward any segmentation you want, identifying risk all the way down to one person"
However, Anderson acknowledged that this specific, granular use of data could create some predicaments for insurers in regard to how it adjusts its premiums and specifically how to handle personal information.
According to Smith, this information will be incredibly useful for clients, as he notes that Milliman's clients want as much information as possible. "Any information that is publically available, they want us to use. No one has said 'no' to anything. In this regard, it's preferable to look at data in its rawest form because it has the most signal," he said.
Smith reiterated that from Milliman's perspective, "We don't think of data as being good or bad; we see it as useful or useless. We say to our clients that if the results were perfect—and they never will be—but you don't use them, then it's useless."
Buluswar agreed, "Waiting for perfection is an exercise in futility. Work with what you have today, make the most of it, and then work toward enhancing your systems."
But in the insurance industry, in particular, data should be used to enhance human judgment that experienced underwriters bring to the profession. "The challenge in our industry is that data models can often augment human judgment, but they shouldn't replace human judgment," Buluswar explained. "The best outcomes come when you're able to harmonize the art of making decisions with science."
Despite the fact that 90% of the world's data has been generated in the last two years, it will only continue to grow. "A friend of mine once told me that if the amount of data in the world today is the size of a golf ball, in four years it will be the size of the sun," Buluswar said. "I don't know if that's true, but even if it grows to the size of a football field, that's a huge increase."
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.