Very few things matter more to car insurers than assessing the customer’s driving risk. Traditionally, this has been centred around vehicle data and the drivers’ demographics. From around a decade ago, telematics added a new layer to this analysis: the customers’ driving behaviour. Now, Driverly is taking this to another level by mapping the driving context data on customers’ journeys to add a 3rd dimension to the assessment of risk and safety.
1st Dimension - Demographics-based Pricing
Vehicle data and driver demographics still play an important role in car insurers’ underwriting and pricing algorithms. This is why customers need to answer so many questions to get a quote. These questions have been standardised by comparison websites, but the game still isn’t over in demographics-based pricing. Smarter insurers use a number of external databases to enrich and validate their quote data, battle fraud more effectively, and build more sophisticated underwriting and pricing algorithms.
2nd Dimension - Driving Behaviour
Telematics has been around for a decade or so, but the penetration hasn’t picked up beyond circa 5% of the market new business sales. This is due to three main reasons:
Cost: Most telematics solutions in the market are in the form of black box devices, to the extent that ‘telematics’ and ‘black box insurance’ are used interchangeably by many. From professionally installed devices under the bonnet, to OBD (on-board diagnostic) devices and self-installed boxes plugged into the car’s cigarette lighter, all these black box solutions come with a big price tag. It’s not only the device to account for, but also the shipment, installation engineer, and logistics team to organise the whole installation operation.
Reward: Most telematics insurers can assess the customer’s driving behaviour within 2-3 months, but they cannot react to it until renewal. This means that it may be challenging to make ends meet on the first year of the policy.
Customer Appetite: Black box devices are intrusive, and hence not particularly popular among customers.
This means that telematics is typically only an option for high-premium policies.
In Driverly, we use app-based telematics, the most state-of-the-art generation of technology. This means that there’s no black box device involved; we gather driving behaviour data from the customers’ mobile phone via our Driverly App.
This smart use of technology enabled Driverly to introduce a customer proposition that’s based on 100% telematics penetration. In other words, all Driverly customers come with telematics data.
We believe this is a game-changing approach to telematics which sets us apart from other telematics insurers.
3rd Dimension - Driving Context
Most telematics insurers – if not all – focus on driving behaviour and its components: e.g., speed, braking, acceleration, night driving, so on and so forth. In Driverly, we’re taking this to the next level by adding driving context to the equation.
We’ve recently concluded a project with Centre of Excellence in Mobile and Emerging Technologies (CEMET), a team of specialists in University of South Wales. With their support, we augmented the telematics data with two layers of driving context data:
Road risk
Real-time weather risk
The tool includes the whole road network of the UK, with information for every road section, enabling us to conduct our risk assessment at the most granular level.
The road risk layer tries to describe everything about a road section that can influence the risk of accident, using as much relevant information as possible – such as its topographical characteristics, traffic intensity, previous accidents records, and a typification of its surroundings.
Weather is an important factor in the likelihood of an accident happening, so it is also included in the tool. The importance of weather is that it can have one impact in the likelihood of an accident or another depending on the driving style of a person... Ice on the A469 road might not impact much on my father-in-law’s probability of having an accident, because in his driving record there’s no presence of harsh breaking and acceleration events. But his neighbour, who breaks and accelerates more strongly every now and then – despite still having a decent driving score – might be at risk. So, we at Driverly treat these two cases very differently from risk perspective.
And that’s how the tool actually works: the algorithm combines generic, contextual information of the routes driven, with the specific information given by the driving style and other behavioural and demographic factors of the customer and policy. All is then fed into our proprietary pricing and underwriting Decision Engine.
One of the goals of this tool is claims prevention. As we are increasing the relevant descriptors of the risk of accidents happening, we can use them for effective accident prevention that will result in an additional claims-frequency competitive advantage for Driverly.
This addition of the Driving Context layer is completely new in the market, and we believe is crucially important in assessing the driving risk. Driving behaviour data doesn’t offer a complete picture unless augmented with the driving context, and in Driverly we’ve done it. For instance, driving 50mph on a nice, sunny day on a motorway may be perfectly safe... Yet the same speed on a rainy day on a narrow country road will be a shortcut to hospital. That’s the added dimension now offered by Driverly for the most comprehensive view of driving risk and safety on the road.