CEOs Concerned about the Time and Cost of Implementing Predictive Analytics
Panel agrees that actuaries must advocate for the tools by presenting potential return-on-investment and other benefits.
Insurance Networking News, December 14, 2012
Many CEOs hesitate to embrace predictive analytics out of concern over the time and expense of launching the models, according to a panel of high-ranking actuaries and executives at the annual meeting of the Casualty Actuarial Society. The panel presented ideas for educating and encouraging CEOs to press forward with such projects during a presentation entitled “What Executives Need to Know about Predictive Modeling.”
“Executives are becoming infoholics,” said Martin Ellingsworth, president of the ISO Innovative Analytics (IIA) unit at ISO, and actuarial professionals and predictive modelers increasingly are aligned with business leaders to improve insurance operations in marketing, risk assessment, rating, underwriting, claims, agent/customer service, catastrophe analytics, and capital-risk management, to reduce the uncertainty around the carriers’ financial performance, he said.
Stephen Mildenhall, CEO of Aon Benfield Analytics, said insurance executives began using complex computer models to determine their exposures in response to Hurricane Andrew, which ripped through South Florida in August, 1992, inflicting more than $15 billion in insurance losses, far more than thought possible, proving that the old analysis methods did not work.
“The case was made for us by [Hurricanes] Hugo and Andrew after our existing models, like those at many other companies, did not perform adequately,” said Alice Gannon, chief actuary for USAA P&C Insurance Group. Executives at the company were used to heavy quantitative analysis, she said, having come from the military, where mathematical precision is important and mistakes cost lives. “They said, ‘Why aren’t you doing more?’” Gannon said.
Gannon then offered several suggestions for advocating for predictive models:
• Pitch the model at a conceptual level
• Estimate the return on investment, and demonstrate how the project aligns with the executives’ other objectives
• Emphasize the importance of data quality
• Be clear about the goal of the model
• Build on prior successes
• Put together a strong modeling team, and invest to sustain intellectual capital
• Be clear about the risks
Alan Bauer, former president of Progressive Direct and a pioneer in both online insurance sales and using credit scores to rating auto insurance, said a pitch for a new model should show how it will affect retention ratios, consumers and the company’s distribution channels. He also suggested giving executives an idea of what reports the models will generate. “Give examples of how the model will help the company do things faster, better and cheaper,” Bauer said.
Bauer also suggested exploring how models could affect competitors. Progressive was among the first to put a rating engine online that showed Progressive’s rates along with competitors’. That decision helped Progressive in a variety of ways, Bauer explained. The move generated Web traffic, which built awareness of Progressive’s name, and conveyed the sense that Progressive is an honest broker.
Occasionally it sent business to a competitor, which was OK, Bauer said. Supposing the rating engine showed a competitive a policy at $400 and Progressive policy at $2,000 for the same risk, “We just sent our competitor a $1,600 loss if we did our pricing right,” Bauer said.
Bauer predicted that ‘big data’ will allow companies to go beyond today’s predictive models to more advanced cluster pricing, which will provide for more refined expense pricing.
For more information on related topics, visit the following channels:
Add Your Comments...
If you have already registered to Insurance Networking News, please use the form below to login. When completed you will immeditely be directed to post a comment.
You must be registered to post a comment. Click here to register.