Claim Context Enables Smart Pricing

What do you know about claims? The approximate location, when it happened, and a few details that your client provided you. This information might be enough for essential claim management, but there is an unused potential to be unlocked by adding context to claims.

It’s a hot topic in most InsurTech newsletters: the insurance industry needs to change from the traditional annual renewable car insurance policy to advanced usage-based insurance (UBI) products. Most stories focus on enabling tariffs pricing in near-to-real-time and creating more flexible insurance products, but insurance telematics offers much more.

Looking at the total worth of motor insurance claims paid out in Europe, Germany ranked number one with nearly 25 billion euros in 2019. Meaning each day, on average, German motor insurance companies settle claims at an amount of more than 68 million euros. It’s without a doubt that claim/cost ratios are the primary factor. To increase the competitiveness of a tariff, insurance companies aim to improve their client portfolio by increasing the proportion of safe drivers. For that, modeling claim frequency and severity through driving risk and assessing the impact on insurance tariffs are needed.

Revealing accident’s reality

The last few seconds before an accident occurs are most crucial for detecting the causes that led to it. Many people can hardly remember every detail. Moreover, the complicated way of submitting claim details, which in most cases is still a manual and non-digitized process, can lead to missing information. In another aspect, a minority of people might communicate misleading information in bad faith to conduct fraud.

Don’t ask for context: measure it.

Was a driver’s vision impaired because of the dazzling sun, was he/she ignoring give-way or even a stop sign, or was he/she driving much faster than the average traffic flow conditions? Our analytics comprises 19 base risk factors from six different categories, taking into consideration the driving environment. This so-called contextualization is crucial for calculating reliable and objective risk scores. But more than that, it helps to track back what happened before the accident – a significant advantage in processing claims. Our approach empowers insurers to understand better how claims come into being – allowing them to take measures to predict, but more than that, prevent accidents.

How to contextualize claims

A trip can be explained through a sequence of time-stamped locations. Longitude, longitude, speed, and heading are, in principle, enough to see where and when a driver has been traveling the roads.
However, to get meaningful insights about mobility data, the trip data must be put into its particular spatial-temporal context. Adding factors as road conditions, sun angle, speed limits, traffic density, road quality and type, road and lane configuration, or weather conditions to each location allows inferring the driving environment.

What is the influence of claims on insurance pricing, and how can it be improved by contextualizing claims?

Jane has almost twenty years of driving history. She appears to be a safe and sound driver, as her insurance record reveals only two minor responsible accidents. Nevertheless, the hassle of getting the claim submitted and repairs done, was not a pleasant experience. Her claims did only have limited costs associated and settled by her motor insurance compared to other drivers of the insurance’s customer portfolio. Let’s have a look at Jane’s last accident: she was driving early in the morning to her workplace when suddenly a herd of wild boars crossed the street. Unfortunately, she hit one of them, but it was not killed and escaped. The damage on the front of the car was, however, severe. According to the insurance’s procedure, the claim was filed a few hours later, indicating the approximate location of the accident, the approximate time, a description of the damage, and what had happened. The car was repaired for 2500 euros, and the claim was settled.

Process innovation

How would the process look like if Jane’s insurance had put at disposal a digital insurance solution? Jane’s insurance developed a mobile app to offer an easy-to-use solution to manage policies, request additional coverages, and file claims. When the wild boar hit the car, Jane could immediately file a claim by simply following a click-by-click process:

  1. The severity of the accident: Material damage
  2. Kind of accident: Animal involved in the accident
  3. Accident site via the automatic location of the mobile phone
  4. Pictures of the damage
  5. Description of the accident

As a result, a digitally empowered process with only five simple steps, not requiring a lot of effort, and done immediately after the accident has happened.

The impact on the claim manager’s process

Thanks to data augmentation, the insurance claim manager, David, does not only receive Jane’s accident information but a contextualized trip of the last few seconds before the accident was detected by Jane’s mobile phone.

  1. Road information: one-lane rural road with a speed limit of 90 km/h, but possible speed due to road curvature only 72 km/h; road roughness index of 6.2 indicating an older pavement with some damages
  2. Road topology: ascending road with a slope of 4.56%; curvature of the road at the accident location of a low radius of 83 meters
  3. Road signs: animals crossing warning 300 m before the hit point
  4. Traffic information: average speed of 78 km/h; jam factor of 2, indicating there is a low probability of a traffic jam happening
  5. Weather conditions: Sunny, 19 °C, 52% humidity, no wind, high visibility, improbable dazzling sun situation

The accident’s context made David conclude that Jane actually traveled too fast in the curve, with a challenging topology for a safe braking maneuver. She respected the speed limit, but she had little to no time to react when the wild boar crossed the road.

The next step: using the data lake to improve actuarial models

Alex is in charge of setting up the actuarial models. While investigating the past claims during a tariff update process, he detected that Jane’s accident was not a rare coincidence on that type of road. With the help of contextualization and data analytics, he was able to derive the following causality. In rural roads, and in particular, within an animal crossing warning area, the probability of having an accident with a crossing animal is ten times higher than in any other section of a rural road. Also, he realized that customers driving on roads with median roughness higher than 8.0 are making use of road assistance 35% more often than the average customer. Thus, he decided to incorporate into the tariff structure one additional question that will serve as a proxy variable for the product pricing: (i) Which are the five most frequent destinations where you travel by car?. Thanks to this question, the insurer can infer, a priori, the probability of a new customer being exposed to animal crossing areas and bad road conditions.

Key takeaway: Adjusting tariffs and pricing dynamically due to context holds the opportunity to lower cost/claim ratios. Adding more variables about when and where people drive helps dynamize pricing and better cluster client portfolios.



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


Test drive our API Suite for 30 days!
Tell us a bit about yourself, and we’ll get in touch with you in no time.