The quantitative imperative: new analytical techniques require new talent. Companies can choose to leapfrog the competition or be left behind.

In the 1970s, property/casualty insurers embraced mainframe computing, generating mountains of management information. Actuaries and CFOs couldn't get enough, and the foundations of modern analytical management were laid.

Since then, profits have been more consistent because loss reserving is more precise; markets have been expanded because accurate pricing has tamed segments once considered too volatile; and overhead has been reduced by automating underwriting evaluations. Smart carriers have used technology not simply to reduce costs, but to grow and increase underwriting profitability.

So it is with the next generation of analytical tools: predictive models. Companies that add statistical modeling analyses to their pricing and underwriting toolboxes can improve risk selection, find new market segments, and prevent the pileup of unprofitable risks when a segment of policies is underpriced.

Companies that do so quickly can leapfrog established competitors. Companies that fail to respond face a slow demise via adverse risk selection.

Most of the analysis done at the leading personal lines writers today is simply averaging. Pricing models group policy facts into more and finer categories, then average the losses. At some point, the law of large numbers, which encourages averaging, gives way to the principle of diminishing returns.

Predictive models usually involve a regression of multiple characteristics against loss ratio or loss costs. The statistics can be complex, but a regression simply seeks and evaluates cause-and-effect relationships. Often, the causes are many, and the model's power derives from its ability to accumulate facts rather than chop them up.

The best known example of predictive modeling is credit scoring. Unlike most actuarial techniques, credit scoring is an import. Few insurance analysts had the data or the expertise to develop the first credit-scoring models, so they accepted the models from the auto-finance industry. Insurance-specific models quickly followed, but they were simply variations on the loan acceptance models.

Just as the first models came from outside the industry, so too will the early model building talent. Actuaries, underwriters, and product managers are skilled at putting information to work--making the connections between data and business decisions. But few of these analysts have training or experience in model development. The insurance industry needs model builders who can work with the insurance experts.

In the past 10 or so years, most insurers have addressed their talent needs by hiring from other insurance companies. But the statisticians and analysts needed for the next generation of analysis are not to be found within the industry. Insurers must include in their talent development plans the recruitment of skilled analysts with no insurance experience.

Managers are comfortable recruiting based on job titles and company history. Managers and recruiters can distinguish "a casualty production underwriter from an agency multiline writer," including the implicit subtleties: experience, education, work habits, interpersonal skills--all the unwritten requirements that go along with the position description.

All that intuition and experience does not apply to this new search for talent. Hiring companies must be explicit about describing positions and objective about evaluating candidates. Executive recruiters must be diligent about understanding clients' needs and creative about searching for candidates. The Rolodex-spinning headhunter is out: Recruiters must be organizational consultants to their clients, researchers during the search, and patient teachers and enthusiastic salespeople in describing opportunities to candidates who don't know the insurance industry.

In short, insurers need to reach outside the industry to acquire new expertise--an investment in talent. Here's how smart companies can leapfrog the competition:

* Describe opportunities clearly, objectively and with minimal industry jargon.

* Choose executive recruiters that can find and evaluate talent, not just hunt for heads in a database.

* Evaluate candidates based on experience and skill, not position history.

* Recognize that candidates from outside the industry have different backgrounds. Avoid disqualifying talent because of unfamiliar styles.

* Sell the opportunity enthusiastically. Do not assume talented people are dying to come to work in a business they don't understand.

* Team up new talent with insurance experts to put the new analysis to work right away.

Michael Cronin is a product management consultant with Cronin Consulting Services, Inc., Kernersville, N.C. Paul Stulgaitis is the president of Blue Rock Consulting, Portland, Maine. They can be reached at [e-mail].