The future of underwriting lies in predictive modeling
January 21, 2013 by Daniel Pierson
As an industry, we need to embrace new methods of doing business. Change is inevitable, and this change can make us all more efficient, wealthier and better at protecting families.
I love baseball. Not only is it a great American sport, but there are so many comparisons that can be made between baseball and business. For instance, a lifetime batting average of .300 is an achievement. Yet, this means the player was unsuccessful in 7 out of 10 attempts. Similarly, in sales, one will have a lot of strikeouts before achieving one sale.
There are numerous other correlations between baseball and business, but I want to specifically address one: predictive modeling based on statistics. Ever watch the movie “Moneyball”? The 2002 Oakland A’s put together one of the most competitive teams with one of the lowest payrolls. How did they do it? They used mathematical models based on a player’s statistics (batting average, on base percentage, home runs, etc.) to determine which mix of players would play best together. The team became so successful that many other professional teams, such as the New York Mets, New York Yankees, San Diego Padres, St. Louis Cardinals, Boston Red Sox, Washington Nationals, Arizona Diamondbacks and many others have hired people to run statistical models for their ball clubs.
So, what does this have to do with business, or life insurance for that matter? Simply put, businesses will continue to use predictive analytics (or algorithms) to improve profitability, better segment and target customers, improve hiring, and provide better services. Don’t believe me? Just look at Netflix, Amazon, or iTunes. All three of those companies accurately recommend products or songs that fit your taste. How do they do it? That’s right, through predictive analytics. As of late, even Google, Capital One and Progressive, just to name a few, are also turning to predictive modeling to deliver their services.
What is predictive analytics?
In general terms, it’s a predictive tool using an algorithm which predicts a certain result or outcome based on a set of many different independent variables.
So, how is any of this relevant to what an advisor is doing in the life insurance world? Well, have you ever wondered how you could be more profitable? Are you tired of a long, drawn out underwriting process for a small $500,000 face policy? Do you have clients that would buy if it weren’t for such a cumbersome underwriting process, the filling out of applications, completion of exams, and ordering of APSs? Or, have you ever wondered how the industry will expand beyond the less than 50 percent penetration rate and finally provide insurance to the masses?
My belief is that predictive analytics can do all of this, and I am not alone. Carriers have already begun piloting programs and, so far, the results have been very impressive. For those non-smokers between the ages 18-60 with coverage up to $1,000,000, the total time from submission to issue is five to six days. Although approval rates are still below 50 percent, it is expected that as more cases are submitted via this platform and more companies venture into this space, the algorithms will improve and approval rates will increase.Imagine being able to apply and issue life insurance in the same fashion that one applies for an annuity (as a ticket order). Imagine the increased profitability to the entire distribution channel (carrier, general agent and producer) when no exam or APSs need to be ordered. How many more cases could you place if the insured received instant gratification by receiving a policy within days of applying for one? And what about those clients who need and want insurance, but just cannot be bothered with the hassle?
But how is this process any different from simplified or guaranteed issue? I’m glad you asked. The difference is that most simplified or guaranteed issue products limit the face amount to significantly lower levels, restrict product type and typically offer only a STD or smoker rate, whereas predictive modeling can allow for the very best class on the very best products.
OK, so you are thinking that all of this sounds good, right? On the other hand, though, it sounds a bit creepy, and there is always the question of whether or not it is an invasion of privacy. This is where opinions can differ.
I, as well as the current laws, do not see this as an invasion of privacy. The majority of data being used is data that is already utilized in today’s underwriting process. Anywhere from 3,000 to 6,000 data points are taken from disclosed information, such as the application, agent report and non med. Only about 1,000 data points come from traditional external sources, which include MIB, MVR and script check — all of which are already being used — and non-traditional external sources such as Equifax. This information includes vehicle type, occupation, housing (rent/own), hobbies (gambling, fishing, hunting), lifestyle (donating to charity, sporting events attended), and exercise habits (membership to a gym, subscription to Runner’s World).
Remember, one single factor does not determine the outcome, but a combination of thousands of different data points that are properly analyzed, can.
Whether or not you agree with the use of predictive modeling in life insurance is irrelevant. If the models prove successful (and they are looking good so far), this will be an evolutionary change in the industry. Carriers will move forward with or without distribution. As an industry, we have all failed in growing the market and penetrating the more than 50 percent that do not own life insurance. Someone is going to figure out how to grab this market share. After all, how much information does Facebook have on their users? Don’t you think it is remotely possible that Facebook and other social media sites could use all the data their users provide to feel comfortable enough to offer life insurance either by creating their own captive agency or partnering with an existing carrier? In fact, we are already seeing carriers becoming more frustrated with distribution and going direct to consumers; albeit at the current time, these consumers are ones a typical agent would not prospect.
As an industry, we need to embrace new methods of doing business. Change is inevitable, and this change can make us all more efficient, wealthier and better at protecting families.
