Manny Rios, SVP of P&C underwriting for USAA (San Antonio, Texas), shared an anecdote at the outset of his presentation at Insurance & Technology's Executive Summit in Carefree, Ariz., that demonstrated the business value of predictive analytics.
Rios said on his flight to Phoenix that he sat next to a man named Arch, a 47-year member of USAA, who had a 1967 Corvette that was not adequately insured. He lamented that even though USAA has a large amount of data on the customer, the company was not aware of the vehicle.
"After 47 years, we should know him," Rios said.
Rios said that though some insurers still employ traditionally reactive strategies for things like customer retention. However, using predictive modeling, insurers can be more proactive when it comes to changes in the consumer or marketplace.
He used a decision tree as an example of predictive modeling. The tree is structured as a sequence of questions with answers tracing a downward path. The original segment (root node) contains the entire data set. Decisions are made for each of the final nodes (terminal nodes) or leaves, based on the observations within each leaf.
Rios used miles driven as an example case for the decision tree. The tree helps separate those who actually have low miles driven per year from customers of prime working age, who often don't realize how far they really drive.
"Everybody underestimates miles driven, so you've got thousands of cars rated less than 10,000 miles," he noted. "We don’t do anything anymore without hardcore modeling and a decision tree."
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