Through various telematics devices, companies can collect data about how drivers behave behind the wheel. The data helps to understand how often and how harshly they break or accelerate, how fast they drive and how this speed compares with other road users. Putting all this together into a single score helps to measure how well an individual drives and it is this information that the insurer can use to provide additional discounts or other incentives to drivers who perform well. In our interview, Dr. Sam Chapman, Chief Innovation Officer and Co-Founder of a leading Telematics Organization explains how telematics data can support accurate insurance pricing and how it can benefit drivers and the environment.
How does “accurate” insurance differ from traditional insurance?
Traditional insurance has historically utilized a wide range of proxies to best estimate risk for individuals based on partial or indirect information. These factors used by insurers are proxies of true risk indicators, established when better data was not available.
Typically these include: vehicle make or model, credit history, claims history, home postcode, age, profession and – until recently – gender (the EU gender directive has already restricted the use of this unfair proxy risk factor on 21st December 2012, other unfair factors will likely follow).
Accurate insurance, instead of estimating risk from unfair proxies, aims to determine the true underlying risk being insured to more accurately price insurance from measuring the insured activity itself. In the case of motor insurance this includes a close understanding of how individuals actually drive, and is monitored by various movement measuring devices. Typically, this approach is known as telematics – the technology of sending, receiving and storing information via telecommunication devices.
Telematics often looks into the variables of an individual’s driving habits such as braking, acceleration, speed and mobile use, to analyze and understand driving behavior from which scores can be produced for insurers.
What will vehicle insurance (pricing) be based on in the future? What are the considered risk factors?
Pricing of motor (and other) insurance is gravitating – as it always has – towards the best known indicators of risk. Currently, the best indicators of driving risk originate from precise information about how people actually drive, for example; shifting from traditional insurance towards accurate insurance. This is of course not highly surprising as the activity being insured is now directly examined thus increased insight into actual driving logically helps prediction for real world risk of driving.
The understandings of driving behavior in telematics insurance are currently distilled due to vast amounts of data into proven behavioral scores or activity events for each individual journey – each of which can be used both by an insurer to help set pricing, but also by the driver to help understand, and minimize more risky behaviors. These are based upon extensive evidence including proven correlations to claims outcomes and also help to give educative feedback to individuals to challenge risky behavior.
As a relatively young science, models of driver understanding are still constantly improving. This is particularly so as they take on better quality data and new analytics. Currently, accurate driver behavior is understood by using either, a smartphone, a black box, an OBD (On board Diagnostic) device or OEM (original equipment manufacturer – in car electronics) to track mobility (typically GPS and accelerometer).
Future risk factors are likely to evolve beyond these systems however to include advanced mobility data such as radar and LIDAR (Light Detection And Ranging) style data to come up with a more complete understanding of driver risk incorporating surrounding driver interactions and proximities.
What can be measured and what can’t (for example, sudden lane changes)? Can it be fair?
Ultimately telematics is currently measured from raw sensors either embedded in a vehicle, added aftermarket or utilized from within personal sensors carried with the driver i.e. smartphones. In all such approaches the raw sensors of GPS and accelerometers are subject to a small degree of variability meaning they must be analyzed to build a complete and accurate picture of each vehicle in motion. The better the understanding of accurate motion the better the capability of the system.
What different types of accurate motor insurance are there? Which one do you think will prevail?
Regardless of the data gathered Usage Based Insurance (UBI) comes in many flavours. For example:
• TBYB – Try Before You Buy, where driver understanding is decoupled from a policy, but is used before a policy application to attract a favourable price based from sample driving.
• PAYD – Pay As You Drive, where policy costs are based upon usage levels, whereby usage is tracked to understand driving mileage or driving minutes to give a dynamic pricing based upon usage levels matching the insurers risk exposure.
• PWhenYD – Pay When You Drive, where policy costs prevent or encourage use in safer times and curtail long late night driving when risk is greatest. These are often sold as Curfew policies.
• PWhereYD – Pay Where You Drive, virtual geographical boundaries (geofence’s) can be used to constrain driving geographically whereby policy risk is minimized by encouraging driving within limited regions. This can also operate directly depending upon accumulating locational risk as drivers traverse roads of differing risk levels.
• PHYD – Pay How You Drive, this aims to directly understand how an individual drives rating according to precise behaviour of the driver.
For each of the above base forms of how telematics insurance can be administered and presented to a user where each approach has particular strengths and weaknesses, for example PWhenYD is proving to be generally less popular in younger drivers as it constrains end users and at times creates risk by encouraging drivers to speed to try and meet cut off times.
The truth that is emerging however, is that not just one approach will prevail. Instead, the good aspects from each approach are slowly being combined to obtain optimised solutions. Nevertheless, the best solution will differ from one country to another, insurers preference or demographic to match the needs of the insurer and the end drivers. Understanding the best approach is another science in itself.
Why do we need a different vehicle insurance model?
Traditional insurance is inherently unfair. Using proxies means drivers are often unfairly rated, for example paying for a demographic bias, or the prior claims of neighbors rather than actually related to their personal good driving. Accurate insurance models are more predictive in that they relate good driving to pricing. Flexible and tailored solutions maybe the future route to success.
How is the environment affected (for example, social benefits) by accurate insurance pricing models?
Telematics not only gives advantages to drivers with cheaper policies, discounts, added value services, but also to insurers with improved risk understanding and better drivers. However wider society also benefits from telematics, for example drivers are educated to drive differently using actual driving data to feedback minor changes.
These small changes encourage smoother driving and decrease aggressive accelerations that reduce risk, but also reduce pollution by optimizing fuel usage. What is more, anonymous data with user consent can be used to understand road pollution, improve traffic management and help to power smart cities of the future.
What are the most important advantages of an accurate insurance pricing?
Drivers policy prices without accurate insurance pricing have to factor into pricing underlying risks from potential bad drivers and fraudulent ‘cash for crash’ criminals. Without accurate insurance, data insurers find it hard to disambiguate risky policyholders from traditional rating factors alone.
Telematics offers motor insurance a measure of how individuals drive, thus, acting as a direct means to understand an individual’s risk as opposed to using discriminatory socio-demographic data. For individuals this means a fair price rather than paying for the errors of others with a similar car or postcode. As well as this the closer observation allows end users to also improve their driving and become safer road users.
What are the biggest challenges on the way to implement them?
Data size and complexity are a huge challenge compared to traditional insurance approaches. To understand accurate risk, it is essential to understand vehicle usage in a fine degree of sub-second detail, this alone presents a huge challenge resulting in 100’s of millions of miles of journey data in sub second precision. To conduct accurate and meaningful analytics with a proven correlation to claims is a very complex task.
Do you think telematics supported Usage Based Insurance (UBI) will benefit insurers, drivers and the environment? If yes or no, tell us your reasons for the choice in the comment section.
Please note that this article expresses the opinions of the author and does not reflect the views of Move Forward.