Although predictive analytics can be applied to a variety of marketing needs, its application in setting prices stands out.
Insurance companies must set prices for corporate customers based on expected rates of enrollee illness, but they have relatively little flexibility to change those prices once they are set. "Once we set a rate, we're locked in," said Carol McCall, VP-research and development at insurer Humana Inc. "And many things can change between the time I set it and when a treatment is paid for."
McCall realized that the industry's traditional actuarial approach, which calculates insurance risks and premiums, didn't work well for her. Specifically, actuarial averages don't help in "creating new products and engaging people in the management of their individual health care," McCall said. "We needed a different technique," she said.
Three years ago, McCall and her team built a predictive engine called Smart Start that determines a tipping point for the company's services, the prices at which corporate enrollees elect different insurance offerings.
A second advantage of Smart Start, McCall said, is its scalability. If Humana were to use its standard model of care for individuals, it would need greater number-crunching of enrollee needs and proclivities.
"You couldn't hire enough actuaries to get that done," she said. "We needed a new math."
McCall received input on her programs from Tom Davenport, professor of information technology and management at Babson College, who said Humana's approach, in a broad way, is valuable for any business.
Using predictive analytics to set pricing and product models that corporate clients will respond to "tells you which companies are most likely to be good customers," Davenport said.
In addition to fine-tuning pricing, the predictive engine also proved to be an aid in product development, allowing Humana to predict which corporate enrollees were at risk for certain illnesses—a key both to product development and intervention.
The program, called Personal Nurse, uses actual nurses who work with individual enrollees to help them make behavioral changes to positively affect their health.
In addition, McCall has developed predictive models around particular diseases, such as obesity, congestive heart failure and diabetes, identifying at-risk enrollees and creating intervention programs.
In other businesses, these might be identified as pain points that drive up costs. At Humana, the results have been impressive, McCall said.
"We measure the impact of this in a couple of ways, but the financial one is in savings," she said. In terms of health care cost reductions, three years after implementing the programs, the savings run into the "multimillions," McCall said.
"We also take those savings into account in pricing, offering richer benefits, or lower premiums or building other services," she said.