Free Preview

This is a preview of some of our exclusive, member only content. If you enjoy this article, please consider becoming a member.

Let me tell you about a burr that’s been under my saddle for way too long.

Simply put, I think that so much of the way analytics and AI are being promoted to the insurance industry is similar to the mountebanks’ magic elixir pitches from the traveling carnival side-shows of old.

Executive Summary

Despite an abundance of optimism over analytics, AI and more, the insurance industry still has challenges realizing the potential benefits of these new insights. Upon simple examination, the reasons are rather straightforward. “Actionable insights” are only valuable when applied to the day-to-day business of insurance and its processes and decision-making.

I consider analytics to include business intelligence, predictive analytics all the way up to artificial intelligence and machine learning, deep learning and more.

I do see considerable value in analytics and new insights drawn from data of all sorts.

First, I am speaking from the perspective of an insurance business professional with 30-plus years as an underwriter, branch manager, broker and home office professional; not so much from the additional 17-plus years of observation I had as an insurance industry lead for three international technology companies, all with heavy focus on analytics. While that experience impacts my vision, too, for now I am taking the line-of-business point of view.

To continue my argument, you are likely familiar with the famous quote often mis-attributed to Albert Einstein that defines insanity as “doing the same thing over and over but expecting different results.” That’s exactly the point I want to make about analytics in insurance today.

There is absolutely nothing wrong with analytic efforts, the tools that are available or the insights that are being developed by many insurance companies. What’s missing in nearly every pitch, conversation, article and way too many corporate analytics programs is the commitment to the most important part of the equation: decision management and process redesign. That’s what’s insane.

I espouse this view: “Insight is worthless until it is acted upon.”

The analytics insights being developed through most insurance projects are simply not being utilized in an effective, consistent and thus reliable and measurable way. That’s because the insights afforded aren’t really augmenting the insurance decision-making processes.

If the business unit—whether it be underwriting, pricing, claims, sales, marketing, distribution management—isn’t involved in deciding what aspects of its decision-making routines might benefit from analytics insights or automation, then how can the analytics efforts be anything but a hit-and-miss effort?
What value does a risk or severity score provide if it is not affecting the business decision-making process around consideration, pricing, or terms and conditions? The same goes for buyer propensity, customer segmentation, fraud scoring, claims severity and next-best-offer. What difference does that insight make if it doesn’t directly affect the decisions being made by insurance professionals and the operations systems supporting them? Also, how can the value of that analytic insight accurately be measured if the results are not being reliably measured because it isn’t being consistently employed?

Here’s an even bigger issue when it comes to the value of analytics efforts: If the business unit—whether it be underwriting, pricing, claims, sales, marketing, distribution management—isn’t involved in deciding what aspects of its decision-making routines might benefit from analytics insights or automation, then how can the analytics efforts be anything but a hit-and-miss effort?

Point 1: Decision-making should be treated as a core business competency that is well understood and consistently applied, with the result accurately measured and assessed.

Point 2: Analytics competency increases its value when there is tight integration between business lines and analysts.

Point 3: It’s when analytic insights are applied to the decision-making process on a consistent basis that the science of analytics works.

There’s no such thing as magic when it comes to business intelligence, predictive analytics, artificial intelligence and machine learning, but there are countless examples of where analytics has changed the game for business decision-making.

Look at how Jim Simon changed the business of investing with Renaissance Technologies. Renaissance uses computer-based analytics models to predict price changes in financial instruments, returning market-leading results.

There’s the great example of Billy Beane, general manager of the Oakland A’s, bringing analytics to baseball. Rather than relying on scouts’ experience and intuition, the Athletics used sabermetrics (the application of statistical analysis to baseball records) for selecting players while ignoring their perceived “weaknesses.” This first phase of analytics use was for strategic decision-making: player selection. Soon, analytics were “operationalized” for calling pitch selection and shifts in the outfield defense.

Consider the Cleveland Clinic and its success in medical performance management, which began with analyzing the telephone wait times of patients and striving to improve service. Applying these analytic insights to the call-handling improved customer satisfaction and medical results.

Capital One in consumer finance and Marriott in travel accommodations are two other classic examples.

Common to every one of these firms and the countless organizations that followed their examples was that analytic insights became a common core to their way of doing business and the making of day-to-day transactional business decisions.

Related articles

These previous CM articles are valuable references to the opportunities and values of analytics and new insights.

Sadly, there are many examples in insurance where brilliant insights have failed to be effectively applied to the traditional business model and thus not provided any or very limited value.

Also, insights aren’t always perfect, and conditions change over time. That’s why analytics-based decision-making is a business capability and not a once-and-done exercise.

My advice is to take some time to explore your decision-making processes; connect your analytics efforts to the business of insurance and find line-of-business executives committed to examine and improve their department’s performance.

Design thinking and decision modeling workshops are a great means by which to explore the decisions that your organization is making today and those that it may be overlooking, as well as documenting and evolving the decision-making processes—both human and automated.

There is tremendous value in analytics-based decision-making. Focus on making this a core competency at the line-of-business level.