5 Myths About Insurance Telematics in 6 Minutes

5 Myths about Insurance Telematics in 6 Minutes

By Chris Carver, ATG Founder and President

There are more… but here are five common myths about Insurance Telematics that should be understood before deciding how best to deploy a UBI solution for either commercial fleets or for personal vehicles used in every day commerce. Failing to take six minutes to read them and little longer to understand them, could doom your UBI initiative.

Myth #1: Insurance-Telematics Is About Driving Data

Nope, it’s about people, the decisions they make during a drive, either a commute or in the operation of their business. The risks associated with those decisions ultimately drive the loss costs. Data is merely a means to understand choices and behaviors that caused the losses. For example, if the business decides to purchase a heavy-duty vehicle to haul higher tonnage loads or opts to drive on a more densely traveled route to deliver its products, those decisions (which have nothing to do with driving performance) can and should be accounted for in a driven risk analysis. This simple principle emerges over and over as a common misconception of what constitutes "driving data".

Myth #2: UBI Data Principally Helps Drivers Lower Rates

We all believe we’re good drivers. Not everyone deserves an immediately lower rate. But, everyone benefits by understanding UBI data and how to earn a lower rate. Risk and Safety Managers certainly need to see the trend reports and analysis that telematics data makes possible. But so too, do vehicle owners so they may better understand how operational choices (types of vehicles and routes, hours of operation, radius of operation, etc.) affect risk. And today's generators of all that driving data don’t tell a full story, the health and well-being of the vehicle operators themselves must be considered. In addition, when drivers are made aware of the data analysis they can better understand their decisions surrounding a trip, as well as how their driving choices impact risk and safety – they must be aware of and understand behavior to change it.

Myth #3: Driving Data Equals Objective Truth

Nope. Telematics data has both signal bias and algorithmic bias. The signal bias is obvious when we look at estimations of the impact of speed on severity. Telematics shows us that severity of a crash is highly dependent on class of vehicle and speed but it tells us nothing about causative factors, such as fatigue and weather. In the accident this weekend on Interstate 78 initial indications suggest fatigue played a role in the secondary collision, as the snow squall descended. Yet, fatigue measured and found in 7% of all driver logs will not likely be found in the data of telematics devices of those involved. Additionally, in other circumstances, we can look at cell phone usage and see that cell phone users are less likely to speed (presumably because their focus is elsewhere). Should we therefore infer that distracted drivers have lower severity crashes? Not necessarily; they would not likely have responded fast enough in rapidly changing weather conditions such as these. Their distraction might have prevented them from braking in time (if at all).

Clearly there exists an algorithmic bias in harsh braking event calculations, which does not take into account vehicle weight. This bias will prevent accounts with larger trucks from ever earning good driver points as by the nature of their heavier loads their braking events will tend to be harsher. Furthermore, event data clearly shows that momentum (p=mass x velocity) is also not constant during a severe deceleration. Crash experts can attest that impulse (the change in velocity due to the force applied at the point of impact) varies based on the initial speed of impact.

Thus contextual data about traffic and weather is clearly needed to establish a more objective analysis and remove both signal and algorithmic bias. Most importantly, we cannot assume that data will yield absolute or objective frequency data; the results are more dynamic than that and our models must allow us to continually revise the predicted frequency.

Myth #4: Telematics Big Data Sets Overcome Bias

Big data vs. thick data… What is meant by “big data” in the UBI world? We like the definition “a collection of data from traditional and digital sources inside and outside the vehicle that represents a source for ongoing discovery and analysis”. Thick Data provides the experience and knowledge from a complex history of primary and secondary research that helps gather granular, specific knowledge about driving safety and a driver’s personal preferences. In other words ... telematics data alone is not a silver bullet for the UBI audience.

Only with both, Big Data and Thick Data, can we understand customer behavior, analyze their choices and account for the selection of quality vehicle safety systems in an insurance pricing strategy which rewards consumer preferences, and leads the game. Driving data differs dramatically when people are aware that data is being collected.

Finally, and importantly, UBI data (big and thick) actually puts the Agent or Risk Management professional in the driver’s seat when selecting the right UBI program for the insured. Carriers have preferences in the types of individuals or businesses they write; they are biased by their loss experience over time. UBI data varies dramatically from traditional territory data, it tells a much more individualized story about the quality of the risk that the agent is submitting. It will be some time before internet submissions are able to instantly find a price with the precision of an informed agent or broker.

Myth #5: Open Data Platforms Kill Innovation

Open platforms achieve quite the opposite, with a caveat for the insurance carrier. The key to a successful UBI program is data and discovery. Open data platforms allow for rapid iteration, and rapid iteration leads to fast failures leading to greater successes that benefit the consumer… or so says the logic in Silicon Valley. The truth is there’s a lot of carnage in this approach to product development and this continues to challenge many risk averse product teams in the insurance industry. The more data sources they have, the more they can access deeper and richer insights for risk analysis, but yet continue to pull the old levers. With an open platform there is an opportunity to combine new privacy protected data with larger data sets from others in the industry and compare how traditional levers affect UBI results before making a commitment. In other words, open data platforms allow a low cost, rapid iteration process for the discovery of what data is right for an insurance company. It also doesn’t mean that the insurance carrier has to build multiple models and throw one or two away, the way software engineers seem to be throwing away apps. It means the ability to tap into many different data sources quickly. And be prepared, some seemingly useful data will provide no new insights, but fail fast and try different things with your customers. Pair the discovery with customer interviews. Driving analytics and usability testing are both essential.

So what’s the point of this 6-minute myth-busting exercise?

First, we must use Insurance-Telematics driving data with empathy -- the inherent biases and lack of context (operational demands, as well as vehicular and environmental factors) can skew data and mask the true nature of the risk. Second, we must be willing to work with the professionals in our industry and multiple data sources to build a broader and more informed reference base. And, third, we must apply the risk insights across all drivers and make the insurance value proposition more meaningful at an individual level regardless whether they qualify for a credit today, or not. That means everyone is made aware of the risks and trends that an effective UBI program can reveal.