Big data and insurance

Big data. Surely you've heard this little two-word catchphrase over the past few months.

Companies in all industries are seeking out new ways to gather, interpret and leverage the massive and ever-growing collection of personal information now available about consumers, thanks largely to smartphones, GPS and social media.

I first became intrigued about the power and potential of big data while reading the 2003 Michael Lewis bestseller "Moneyball: The Art of Winning an Unfair Game." The book examined how General Manager Billy Beane's Oakland A's mined statistics to identify undervalued players, allowing the small-market franchise to successfully compete against big-market clubs with budgets several times bigger than their own.

"Moneyball" asserted that the way baseball organizations — scouts, coaches, managers and the front office — evaluated talent over the past century was subjective and often flawed. Too much emphasis was given to traditional statistics like batting average, runs batted in and stolen bases. The A's used sabermetrics — specialized analysis of baseball through objective evidence, especially statistics that measure in-game activity — that demonstrated on-base percentage and slugging percentage are better indicators of offensive success.

Using this revolutionary approach, the A's, with a team payroll of $41 million compared to $125 million for the New York Yankees, made the playoffs in 2002 and again in 2003.

This is a great example of an organization taking advantage of available (big) data to make better decisions that significantly enhanced organizational success.

The insurance industry has begun to use similar principles in very tangible ways. What immediately leaps to mind is Progressive's patented, proprietary Snapshot. Some of the company's auto insurance policyholders voluntarily plug the little device into their cars and it measures their driving activity, like how often they slam on the brakes, miles driven, and how often the car is driven in the more dangerous time between midnight and 4 a.m. It doesn't care how fast or where people drive.

But the "better" users drive, the more they can save — up to 30 percent compared to their original rate (which can't go up because of Snapshot).

See also: How Progressive's Snapshot works

As Chris Stehno, senior manager, Deloitte Consulting, mentioned in a recent session about predictive analytics at the LIMRA Life Insurance Conference, auto insurers had basically tapped out existing data sources about policyholders, so they came up with this data creator that provides additional, very accurate data that can help identify and reward "good" drivers. I would think this kind of real-world data would be a lot more helpful than figuring in a driver's credit score.

Next page: Predictive analytics in life insurance

Stehno also mentioned that smartphones are becoming a consumer's Snapshot for other miners of data. A smartphone provides a ton of information about its user and his or her lifestyle, and much of that information is being collected and analyzed, whether consumers realize it or not.

This was happening years ago (and still is) with web-browsing habits. Think about the last time you searched airfares online for a possible trip to Chicago or browsed sites for research on new HDTVs. Funny how the ads that pop up in your next browsing session tend to be about Chicago airfare and hotel deals or a new Panasonic flat screen.

Think also about the apps you use on your smartphone. The company behind that app knows how you are using it, and you can bet they are working on ways to leverage that information to entice you to buy or consume more of their product or service.

But let's fast-forward to life insurance and how predictive analytics are helping insurance companies tailor their marketing. Companies used to gather information about who might be likely to buy life insurance. Stehno mentioned in his LIMRA session that now companies — using big data — not only know who's likely to buy, but who's likely to qualify. How helpful is this? Perhaps a new product is made "available" to everyone, but the company only markets it to the group it really wants to target — a specific demographic known to be made up of a high percentage of people who would qualify for the coverage. By combining who's likely to buy with who's likely to qualify, you create efficiencies. If a person is likely to buy, but you know he won't qualify, why bother spending any money marketing to that person?

It can also help with the upsell/cross-sell issue, Stehno notes. Suppose a P&C agent sells auto insurance to a customer and is looking to cross-sell life insurance to him as well. If the agent approaches him about it and he doesn't qualify, the agent has likely made that auto insurance customer angry. Predictive analytics can tell the agent if the person probably won't qualify before the agent tries to cross-sell him life insurance. On the flipside, if the analytics tell say the person is likely to qualify, that agent has a hot prospect.

Savvy producers have been gleaning information about what's happening in their clients' lives for years now by following them on social media sites like Facebook, LinkedIn or Twitter. People post information all the time about major events — expecting a child, moving, a new job, etc. Knowing these types of things in a timely manner creates sales opportunities.

We are still just scratching the surface of the potential of big data and predictive analytics in the life insurance industry. Stenho's session at the LIMRA conference was titled, "Predictive Analytics: Here to Stay or Passing Fad?"

Here to stay. It will become the norm.

For more from Brian Anderson, see:

Think Pink: Sales psychology lessons for the life industry

Great opening questions for DI producers

What MetLife's learning from its Walmart pilot program