INTRO: The use of big data is most effective when harnessed to an overriding strategic vision. Howard Mills, chief adviser for the insurance industry group at Deloitte L.L.P., explains how the use of big data to turn out the vote in the 2012 U.S. presidential election is instructive for the insurance industry.

Politics and insurance are two of the world's oldest industries, and have shared a certain interrelationship throughout history.

One can imagine that when the first written insurance policy appeared as part of King Hammurabi's code in Babylon sparing debtors in the event of catastrophe, that the king's political advisers saw this as a good thing for his popularity rating. Who knows? At that moment, perhaps a regional governor somewhere might have said state, not federal, regulation would have been the better route. And surely his legal advisers were concerned about contract certainty — who would define a covered catastrophe, for example.

But it worked, because almost 4,000 years later, both politics and insurance are still central to modern life. Both, in a sense, are about managing risk. Politics is about convincing your target market to centralize risk management with you — as president of the United States, for example — as the ultimate risk manager, while insurance is about convincing your target market to share the risks we face in order to manage the individual impact of that risk.

Practitioners in both industries face similar challenges: increasing market share in a seemingly mature market. So with January's presidential inauguration a recent memory, it may be useful for insurers to look back at the lessons of the 2012 presidential election.

%%BREAK%%

One of the biggest messages, if the pundits are to be believed, is that intensive use of data is a necessary tool. But data alone is not enough. Both campaigns had access to the same data, yet both came up with remarkably different readings. In the losing campaign, CBS News reported, one adviser described the candidate as “shell shocked” by the loss.

Why? They looked at the data and made the wrong assumptions. As Jan Crawford of CBS News put it: “Those assumptions drove their campaign strategy … Those assessments were wrong.”

Let's pause before we succumb to hubris and say we could have or they should have done better. The truth is they based their assumptions on what was historically correct, much as the insurance industry does. Unfortunately, as the industry learned with its initial pricing assumptions for long-term care products, past performance may not adequately predict future results.

Does that mean an industry based on data, on the analysis of past performance and its use to predict future results, no longer can rely on that data?

No, absolutely not. The winning campaign in the last presidential election relied even more heavily on big data, according to Time magazine. Everything was data-driven, and the data was tested, retested and monitored obsessively. But it was subject to a certain vision — to leave no potential supporter at home on Election Day.

%%BREAK%%

To that end, the campaign developed a “persuadabilty score.” This score, Lois Beckett reported for ProPublica, “tried to capture not just a voter's current opinion, but how that individual opinion was likely to change after interactions with the campaign … To pinpoint voters who might actually change their minds, the Obama campaign conducted randomized experiments.

We will never know how well this worked, but the final result showed an electorate significantly more skewed to the winning party than to the losing campaign — and more than many outside observers expected.

Ours is an industry that needs to be able to find persuadables.

Perhaps, as did the winning campaign, our industry should not allow itself to be held hostage to history or the shared assumption of a mature market, but should work on innovations, whether in product development or distribution, that encourage the market to expand.

For that, data is an essential tool, properly harnessed to a vision of what should and could be done.

Howard Mills is chief adviser for the insurance industry group at Deloitte L.L.P. and a former superintendent of the New York Insurance Department. He can be reached at 212-436-6752 or at howmills@deloitte.com.