The massive proliferation of easily accessible data combined with the increased power of modern analytical tools has the potential to transform the insurance industry dramatically over the next decade. The strategy and operations of insurers in the near future could be nearly unrecognizable to current market leaders.
Insurance has always been an industry driven by data and analytics. Early insurers outperformed their peers by gathering more information faster and evaluating it more effectively to pool and price risk better.
In order to gather data, insurers adopted critical tools back in the analog or pre-digital age. They employed large networks of people to find and interview prospective risks, then they created paper forms so the information those people provided would be sent to the home office in a standardized format. Then they hired intelligent people to process that information and make a decision.
Many insurers essentially operate the same way today. The network may be commissioned agents instead of employees; the paper forms may be electronic and the smart people in the home office have better information storage and analysis tools at their disposal, but the underlying assumptions are basically the same.
Three of these underlying assumptions are:
1. The best way to find out about prospective insureds is to ask them for information.
2. The right time to make decisions about a prospective risk is after asking them for information.
3. The best models are based on the information contained in the questionnaires insurers have historically used.
The current data and analytics environment call these three assumptions into question.
In a world where data aggregators such as Acxiom, Experian, LexisNexis, Dun & Bradstreet, GoogleEarth and dozens of others, have already answered any question an insurer might have about a prospective risk, it's possible to underwrite many products without asking a single question. The answers are already out there, in structured electronic form, for sale in the infosphere.
This means that it's now possible for insurers to know enough to underwrite a risk before they engage directly with the prospective insured. This could flip the standard practice, in which marketing shovels applications into the top of the funnel and underwriting throws most of them out. With better advance knowledge about prospective insureds, insurers can start conversations only with the risks they actually want and can price them accurately at the beginning of the conversation rather than at the end.
When the cost of data acquisition is dramatically reduced, insurers can shift from the blue whale method of business acquisition-taking a big mouthful of seawater to capture a tiny handful of krill-to a dolphin model, grabbing just the fish they want.
In a world of ubiquitous data, and with the tools to analyze it effectively, much of risk selection can move from underwriting to marketing. This has the additional potential benefit of broadening the range of data sources insurers can use in their modeling. States may restrict the data sources permissible for use in underwriting, but no state currently restricts the data usable for marketing in the same way.
In addition, the improved analytic capabilities of modern modeling tools offer actuaries and underwriters an opportunity to rethink some of their models, some of which have evolved organically from the analog age and have never been questioned systematically. Which constellation of data elements really are predictive of loss and profitability? What is the marginal value of any particular data element? How does that compare to the cost of gathering the data? Some insurers took steps in this direction years ago by applying predictive models to the need to order expensive motor-vehicle records, and only ordering them additional information could make a difference to the outcome. What if every data element were subjected to the same level of scrutiny?
An insurer designed to thrive in a modern, digital infosphere might be structured very differently than one that evolved from the analog age. For example:
* Customer acquisition could be expanded and marketing given more responsibility for risk selection.
* Underwriting could be streamlined to focus mostly on modeling and handle a very small number of special cases.
* The customers' purchasing experience could be transformed from one resembling an audit experience to one more like applying for a credit card.
* Agents would do only what they're really paid for: selling; and wouldn't be burdened by the need to gather information.
We are only at the dawn of the age of ubiquitous data and powerful analytics. While technology investments and strategies are necessary, the slow evolution of analog-age business structures will be the greatest challenge for most insurers.
INNsight is exclusive commentary from Novarica. Matthew Josefowicz is partner and managing director at New York-based Novarica, a research firm focused on markets, operations and technology in insurance and financial services.
