Pro-Actively Using Predictive Modeling In Insurance
We live in a time of data. Data predicts how we eat, shop, communicate, watch certain programming, and even how we vote. Companies take that data and then target specific advertising when we are on social media or searching online predicting that we might buy their items based on past searches. We have all experienced this type of marketing on Amazon or Facebook.
A well-known cautionary tale on the use of this type of targeting based on predictive analytics happened back in 2012 in the retail industry, and it provides a valuable lesson for sales and marketers in all industries. Target developed an algorithm to predict when a woman was pregnant based on her purchasing behavior—buying more lotion or vitamins for example. With this knowledge, they would then market to those expectant mothers, coupons in the mail for things like cribs and maternity clothes. In theory, this seems like a great idea for effective marketing. As life changes happen, consumer buying habits change and Target could take advantage of this knowledge to send advertising accordingly.
However, Target’s predictive modeling was almost too good. One father was angered when his teenage daughter was mailed baby-related coupons and went to a Target store to demand an apology. A Target representative called the father to apologize (twice!) about the situation a few days later. The father replied “I had a talk with my daughter. It turns out there’s been some activities in my house I have not been completely aware of. She’s due in August. I owe you an apology.”
The moral in this story is “don’t be creepy.” Target could have mixed other coupons in with this to avoid the issue altogether. There is a balance of using data to be strategic while not making your target audience feel like their privacy has been invaded.
In the insurance industry, underwriters will continue to find new ways to use data to predict risk and help provide a more personalized customer experience
What is Predictive Modeling for Insurance?
Predictive modeling uses known results to create, process, and validate a model that can be used to forecast future outcomes. By analyzing historical events, there is a probability that a business might be able to predict what would happen in the future and plan accordingly.
The accuracy of the predictions depends on the data and the algorithm used to make the predictions. In the past, insurance underwriters relied on experience, market knowledge, intuition, and oral history more than data. However, as more analytical tools became available, personal and small commercial carries began to apply the models to automate basic underwriting tasks, risk profiling, and scoring models. Today, predictive modeling has become the norm for the insurance business.
In underwriting, predictive modeling uses a variety of data both internal and external, and statistical techniques to make predictions about underwriting desirability, losses, and/or pricing. The data looks at everything in a potential client’s business category across the country as well as lists the customer’s behavior, needs, and expectations. Predictive modeling analysis allows insurance companies to minimize their risks, but also tailor the policies they write to fit their customer’s needs.
Predictive modeling data can assist insurance companies in fraud detection while getting deeper customer insights allowing them to make more pertinent policy suggestions and ultimately improve customer experience. Because of big data software, insurance companies can save time and money by automating internal processes such as claims or property assessment.
Using Predictive Modeling to Recapture Customers
Underwriters can also use predictive modeling data to create a marketing program for current or past customers. For example, one way to use predictive analytics is to recapture customers (policyholders) you once had. The best way to recapture them will be to make it simple for them to again consider your product. What are the ways to win back a customer using predictive modeling?
• Call-out with all-important API’s to update their information, such as business locations, number of employees, payrolls, etc.
• Use your predictive model to settle on a policy premium price so the decision can be made immediately
• Create email campaigns to deliver a message on why they might want to consider being a return customer
• Use targeted content, such as white papers, infographics and blog posts to attract the customer
• Give them the ability to make the decision “on the spot” and through your company’s site
• Get personal with the policyholders. For a personal touch, follow up with a phone call to encourage them to consider buying now.
Use a variety of marketing techniques to attract customers and avoid being “creepy” like the Target example above. Not everything will work out as you planned so it is important to test and refine your recapture program as you go along.
Predictive Modeling Is Here to Stay
Predictive modeling is present in many aspects of our personal and professional lives and it is not going away. Companies will continue to develop new ways to use the data to predict consumer behavior. In the insurance industry, underwriters will continue to find new ways to use data to predict risk and help provide a more personalized customer experience.