Insurers Need Robust Predictive Modeling to Meet Policyholders' Needs: E&Y

New York-based professional services firm Ernst & Young released its first Global Consumer Insurance Survey yesterday, which gauged how 24,000 consumers in 23 countries feel about the insurance industry.

The results indicate that insurers must be proactive in developing robust data, analytics, and predictive modeling capabilities to deliver optimal customer experience and reduce customer churn, according to David Hollander, E&Y's Global Insurance Advisory Leader. With 52% of respondents saying that they would like to own more than one insurance product from the same carrier, but only 11% reporting that as reality, there's a lot of cross-selling opportunities that can be leveraged.

"If you built a model of what characteristics those 11% have, using both internal and external data, you can better segment in your population people that would buy and what products they would want to buy," Hollander explains. "It's not about pursuing them with a high-pressure sales approach, but having a qualified group of people that matches the profile of those who have already bought multiple products. You can be much more intelligent and granular about what you do."

On the retention side, Hollander notes that about 25% of P&C customers have switched product providers over the past five years. That number could've been reduced by about 6% had companies identified people who were likely to switch and reached out to them before renewal, he adds.

"If they had been approached and contacted about their needs, they were likely not to switch. Can you build models to find people who are likely to consider switching, then have your renewal or conservation activities very targeted at those people instead of a blanket call to everyone based on their renewal date?" he asks. "We think there is a lot of potential to build models based on your own experience, then integrating those models into the real-time system and saying 'Now it's time to reach out, and here's what you want to be discussing with them.'"

Insurers who develop modeling and analytic expertise for marketing, sales and service are building skills that can be applied elsewhere in the business, Hollander adds.

"Clearly you want to have models, for example, based on claims that are likely to be problematic," Hollander says. "It's not one or the other. We have clients who have looked at sales, servicing, claims, and even some back-office operations and see where it makes economic sense to build those models."

[Read how Liberty Mutual Commercial Insurance leverages predictive modeling for claims.]

The insurance industry has made major strides in claims servicing over the past few years, according to Hollander. The level of satisfaction with the claims experience "was much higher" than E&Y expected, he said, noting that a third of people who had a claim rated it a 10 out of 10. Over 70% had an eight out of 10 or higher.

However, this has made it more difficult for insurers to differentiate competitively through claims service. A poor claims experience will almost certainly drive a customer away, but the difference in satisfaction between people with no claims as opposed to people who had made a claim was negligible.

"The investments that carriers have made in claims has clearly paid dividends," he says. "The data certainly says that if you haven't made investments and have a strong claims process, that's certainly important for you. But it also says you can't really differentiate on it. It's now becoming the standard: People expect timely service, communication, and fair payment."

Nathan Golia is senior editor of Insurance & Technology. He joined the publication in 2010 as associate editor and covers all aspects of the nexus between insurance and information technology, including mobility, distribution, core systems, customer interaction, and risk ... View Full Bio