Insurance Pricing and Telematics:
A Historical Contradiction?

This post goes out to everyone in the insurance sector but especially to those who have tried to integrate telematics in their insurance pricing models. 

Why are most insurance companies hesitant to create telematics products? We will make a strong hypothesis now: it’s because nobody really knows how to translate telematics data into their pricing models. If you don’t agree to that directly, give us a chance to convince you and put forward a solution to this problem by the end of the blog post.

The rising of PAYD tariff popularity

With the global Covid-19 pandemic, many customers were evaluating to switch to Pay-As-You-Drive (PAYD) insurance tariffs simply because they did not want to pay for insurance while not using the car. Their vehicle was mostly stationed at its parking lot due to worldwide lockdowns and home office working schemes’ uptake. Recent numbers, however, show that mileage or in the general distance has already almost reached the pre-pandemic level again. Is that the sudden death of the PAYD model’s popularity? We argue no if you do not consider the basic distance-based PAYD tariffs only but also innovative ones, built on a monthly base tariff, to cover risks of the non-moving vehicle and a variable tariff based on the number of days that that vehicle moves. Even if the world will get back to normal at some (hopefully soon) point in time, partially remote work is sure to stay as it turned out to work pretty fine.  

Simple ABC analytics and basic PHYD as the next evolutionary step

However, PAYD is the simplest, although most common telematics-based motor insurance product. More and more insurance companies have tried to innovate and create low-risk customer portfolios with Pay-How-You-Drive products in the past few years. Many companies based their tariffs calculation on simple trackers or mobile apps, collecting trip data, and calculating risk scores based on acceleration, braking, and corning (sometimes speeding, too), commonly known as ABC analytics. The resulting insurance products were disappointing for both consumers and the insurer, especially the actuarial and underwriting experts. What happened with the brilliant idea of including driving behavior and the consequent accident risk into an insurance product resultantly? Consumers who opted-in got, in most cases, only a small discount upfront or a reduction for the following year. The positive reinforcement mechanism of being granted a discount faded over the next period, and most of the time, no long-lasting effect on driving style could have been observed. On the insurer’s side, telematics was mostly seen as a significant integration effort as it was challenging to incorporate into their pricing models. More than that, standard telematics-based technologies do not offer any correlation to accident risk or claims – the primary element for pricing. 

Traditional insurance vs. a data-driven approach

The solution consists of, first, better driving behavior analytics and resulting scoring. Second and most importantly, a translation of risky driving behavior and events to actual claim costs and the impact on insurance tariffs. But let us dive first into how traditional insurance tariffs are calculated. The price you pay to insure your car for a certain coverage depends on the car’s model (car model price and horsepower), e, your driving experience, and some demographic information. Advanced PAYD / PHYD models’ primary goal would be to offer good drivers rewards, limit claim costs due to reduction of risky drivers in the portfolio, or compensate possible losses with adjusted premiums. All that while respecting the insurer’s decision about the impact of telematics on the tariff, in terms of maximum discounts or increases.  

Driving behavior under the microscope

How would that work: it’s as simple as your trial period on a new job. Either you perform, or you don’t perform. Insurance companies have to collect trips over a significant period for an individual driver, enough to derive usual mobility patterns and driving characteristics to extract specific contributory risk factors. Such risk factors could be, for instance, aggressive driving, disobeying priorities, being exposed to low visibility or slippery road conditions, being involved in harsh maneuvers, or exceeding speed limits. These risk factors need to be seen in their context, meaning where the event is observed – rural, urban, or freeways. Each risk factor contributes in each environment differently to the accident risk and eventually claim costs.

The resulting driver profile consists of distance traveled on different road types, the observed risk factors per road type, and the estimated distance traveled per year. Matching these results with recent road accident statistics enables insurers to weigh risk exposure according to severity. This constitutes a major advantage compared to traditional telematics: by using road accident statistics from well-known risk factors, insurers can without any ifs or buts deploy telematics as a plug-in to existing products without having to integrate risk factors individually into their Generalized Linear Models (GLM), eliminating likewise the dependency between claims and telematics data.

So far, so good, but still rather complex to integrate with pricing. Thinking about the practical application of these advanced scoring methodologies, wouldn’t it be great to have a dynamic pricing tool to play around with various insurance product presets like PAYD, PHYD, and even more advanced Pay-Where-You-Drive? Stay tuned to one of our upcoming insurance article blog posts, and we will show you how the future of dynamic pricing looks like in reality.

Be prepared for the connected car era

In the early telematics initiatives, the insurers took the lead in collecting driving data through expensive devices. Most of those initiatives failed. 

The number of global smartphone users has continuously been growing ever since the first smartphones hit the market.  With smartphones in almost every driver’s pocket, there are new opportunities to collect mobility data and create accurate profiles at a low cost. Combining risk factors with road accident statistics to translate them into concrete exposure and severity metrics allows integrating telematics hassle-free into insurance pricing. 

We don’t need to have a fortune-teller ball to forecast that in the (near future) connected vehicle technologies will replace the smartphone in insurance telematics. Car manufacturers are rapidly deploying their connected car platforms and APIs, providing even more accurate driver profiles. Suppose insurers do not seriously take into account data-driven pricing. In that case, car manufacturers will be better positioned to act as risk-takers by bundling insurance into a single monthly mobility budget for the users.

Key Takeaway: Translating telematics data and risk scores into concrete insurance products requires accurate driving behavior profiling and matching it with context and accident statistics. It’s a complex task, but with an entirely data-driven approach, the competition’s main differentiating factor is to offer your best drivers the lowest insurance tariff.  

Share on social media

Share on facebook
Facebook
Share on twitter
Twitter
Share on linkedin
LinkedIn

BOOK YOUR DEMO NOW

We'd love to show you the power of our data-driven products.
Book your personal demonstration now!

Receive the latest news

DID YOU ENJOY READING OUR BLOG ARTICLE?

Subscribe to our newsletter and get notified.