Risk & Analytics

Behavioral Analytics… Predicting how I knew you would say that…

Share this post:

Usage based insurance programs are on the rise; from how a driver behaves behind a wheel to how often or where that person does their driving. Once collected the behavioral information is analyzed and being used to drive greater precision into risk categorization and pricing.

The term ‘telematics’ has come to represent this popular model where a policyholder is given a plug-in device for their car to report back selected data points to the insurer (such as hard braking, quick acceleration, and average speed). Behavioral acts like these, of course, create a profile by which the insurer can start to assess risks associated with the way the particular vehicle is operated. The insurance industry, through a clever marketing scheme, obtains the policyholder’s (or potential policy holder) buy-in to give personal information about their driving habits in exchange for potential discounts. [1] (There are other creative “reward” ideas in this space as well.)

Getting verifiable driving data from a policyholder or prospect has always been a challenge for insurers. Telematics has begun to resolve some of that. This has a direct impact on the ability to segment and pinpoint to an efficient price. Because insurers want to make a profit, the average carrier will build in sufficient cushion on pricing to allow for a misplaced driver. In effect, it accounts  for a pooled segment of drivers having an outlier or two who was not correctly placed. The insurer plans on this and assigns a higher rate. Members of that segment all pay more because of an imprecise pricing capability.

Although popular, telematics is still in its nascent stage, this start is promising, not only because it is aligning policyholder interests with the insurer (sharing information in exchange for a monetary savings), but also because the policyholders are volunteering the information. This is happening in today’s social environment where any perceived or unauthorized use of information is a contentious roadblock to avoid.

(An interesting point is the carrier is allowed to know and use the information because it was directly supplied by the policyholder (via the device). If nothing else, the conversation on when and how personal information is willingly shared becomes a conscious act.)

Through the collection of behavioral actions the data begins to detail the preferences and habits of individuals. Those preferences expand to form segments and, with analysis, the segments and the policyholders’ actual insured driving performance (loss results) eventually contribute to a complete picture that provides an accurate profitability index and leads to pinpoint pricing.

One additional contribute on this front is the behavior of drivers when they know their data is being recorded and shared. I have had any of a number of people tell me that when they put the tool in, they became very conscious of jack-rabbit starts, sudden braking and high speed. This suggests that the telematic tool had an influence on behavior – safer driving?

There are other commercial opportunities in behavioral analytics to be explored. From a marketing focus we can take a look at digital interactions. This other stream is less about driving habits and more about acquiring potential applicants. Digital interaction can be cross-referenced with customer, device, service, network, location and business data to provide a unique but telling snapshot of a policyholder. Behavioral analytics already goes down the road of buying propensity.  However, as insurers open up their multi-modal access points, capturing and analyzing the information about those commerce related events will gain importance. We should expect to see marketing campaigns based on behavior patterns in addition to the consumption of particular insurance products, patience with the portal (abandon point), use of certain electronic tools, and other behaviors.[2]

One usage based calculation that a couple of insurance companies are pursuing is the frequency of use and miles driven. This is an interesting carry-over from older analytics that suggests that the driver who drives less has a lower exposure rate. Reasonable enough to think through. But what about someone who drives less because they take part in a social networking program for ride sharing?

Just recently I was read an interesting article in I&T (Insurers Must Adjust to the Sharing Economy by Nathan Golia).  The article addressed the trend for offering and accepting rides through a social connection in San Francisco and how that may begin to put coverage for such occurrences in question. The concept provides a new slant as a result of shifting societal norms.

And what this may be raising is the classic need for any underwriter to understand the exposures. It is not that an insurance company does not want to cover this change in traditional ridership or responsibility (although the first reaction is to pull back). More to the point is the need to assess the potential risk and exposure. Tell an insurer what their experience (losses) is likely to be and they can adjust the rate.


[1] In case you are wondering, I will have to admit that I have not done this, yet, as my driving habits, with an M3 and 5-series sedan, are not likely to look discount worthy. I will avoid facing that truth for now.

[2] check out  “Oh behave! How behavioral analytics fuels more personalized marketing”


Senior Managing Consultant - Insurance: S&T

More stories

Four ways to stem the growth of financial crime in insurance

Insurers have long understood that the ability to quickly detect fraudulent claims not only stems losses, but also minimizes the impact to legitimate customers as well. But as more complex financial crimes like money laundering, employee fraud, and corruption have gravitated toward insurers more recently, many don’t have the programs in place to combat these […]

Continue reading

How digitalization of life insurance will transform a static industry

Life insurers have a big problem: Younger adults are not buying life insurance. Despite year-over-year population and GDP growth, the number of U.S. residents under 45 who applied for life insurance in 2018 fell 2.7 percent from the previous year, according to the nonprofit group MIB. Insurers have turned to emerging markets such as India […]

Continue reading

How big data, AI, and automation impact insurance

Data is the new gold… Data is the new oil… Pick your metaphor, but for insurers, data is the asset their critical business processes require in order to run effectively and produce actionable results. Data is a critical component of marketing, underwriting, product development, claims, distribution and IT operations, to name a few, and insurers […]

Continue reading