Insurers ignore multi-source data sets for UBI at their peril

As the connected car ecosystem continues to evolve, more sources of data are becoming widely available creating an exponentially more complex landscape for insurers to navigate.

According to Vodafone’s report On the Road Again: The Evolution of Car Insurance, the Association of British Insurers predicted in 2012 that there will be 89M telematics-based insurance policyholders by 2017.  In the US, Vodafone says that five of the 10 US insurers currently offer UBI policies, and similarly to the UK the market, is expected to grow fast over the next few years.

In fact, the Association of Insurance Commissioners claimed in June 2012 that the 20% of the 200M policy market could be UBI-based by 2017. Yet another 2016 article Insurers’ Profiling Will Change As Autonomous Vehicles Hit The Road, Says Expert, claims that insurers will rely “less on the personal traits and behaviours of drivers and focus instead on the technical features of cars when setting car insurance premiums in the future”.

The article’s author, Stephen Appt who works at law firm Pinsent Masons and who specialises in the regulations of autonomous vehicles, explains that car insurers currently analyse a wide range of data, such as the history of a driver’s accidents and with telematics data they assess how safely that person drives. As the move is taken to go beyond simply connected vehicles to autonomous ones, this will change and with it the insurance industry will have to evolve, in his view. They will need, for example, to assess how they assess risk and how they set premiums. Complex software and systems that will dictate how a car drives, as opposed to being driven by a human-beings who will simply become passengers, will require the insurance companies to make their risk assessments on a technical basis.

Data source explosion

Even with just connected cars, the range of data sources is bound to increase. “On average there are five data sources used by the connected car ecosystem,” says Nico Gollwitzer, head of telematics at Vodafone. He adds: “They include but are not limited to acceleration, location, time, type of road, local speed limits, weather, car diagnostics systems and driver ID.”

So, to what extent is the increasing array of data sources increasing complexity for insurers and how can they address it to turn it into a competitive advantage? Gollwitzer finds that the increasing number of data sources is also influenced by country-specific legislation: “Insurers may not be allowed to have access to detailed data sets and only to aggregate driving behaviour KPIs. Some data requires detailed knowledge of the internal car components and how they communicate – this includes, for example, diagnostics data.”

Increasing complexity

Gollwitzer explains that the increasing complexity creates a huge diversion for the insurer. This, ultimately, makes it harder to focus on what it is as an insurance company does best and it makes it more difficult for it to create any real competitive advantage. “In this multi-jurisdiction environment an experienced telematics service provider active in multiple countries, with a strong automotive heritage, would be a valuable partner,” he explains.

TU-Automotive asked him: “How will connected car insurance evolve over the next 5 years and beyond through OEM ascendency and new streams encompassing the wider IoT ecosystem?”
He replied by suggesting that an increasing number of cars will have a factory-installed telematics system already on board. In his view the data is the critical success factor that insurers need to maintain their commercial competitiveness and so they need to have the ability to “tap into this data, which is typically stored in an internal platform of the OEM”.

Gollwitzer then comments: “In addition, new trends of using telematics services in homes make the accessible data sets even more complex. Using that data to keep customers loyal and increasing customer intimacy will be critical to the future success of insurers. The classic insurance risk models will have to be extended to include the cost of customer acquisition.”

Privacy and security

Complexity may also come from the upcoming changings in privacy and security to respect compliance requirements in order to best position new solutions.  He finds that data security and privacy is already the single biggest concern of car makers: “As soon as the international IT security research community will shift their interest from car makers to insurers, it will become a critical brand protection topic also for insurers.”

Therefore, he advises insurers to “carefully review their installed base of devices and infrastructure to protect themselves from loss of confidential data or even access to sensitive customer information in their backend.”  To meet this challenge there is a need to have a robust partner ecosystem in place and there needs to recognition that “security in this environment will need to be consistent with any other large IT programme,” he adds.

Overcome data fragmentation

So with this increasing complexity, how can insurers and other players within the connected car insurance ecosystem overcome data fragmentation by normalising and enriching data from multiple sources to find the key to a successful UBI programme? In Gollwitzer’s view the key to standardisation is the “adoption of cross-industry telematics platforms, which will permit car makers and insurers to share data in a normalised way without having to look at time-consuming standardisation bodies.

To conclude, now is the time to get ready for multi-source data sets. In their AIG case study of March 2014, the Vodafone and Towers Watson confirmed that UK insurance company AIG Europe, part of the international group AIG, had signed up to run a pilot. In it Duncan Anderson, global head of pricing and product management at Towers Watson, offered a word of caution that is probably still worth noting today: “Launching and maintaining successful telematics motor insurance products involves more than simply sourcing a device and considering basic driving events.”

There is also a need to analyse granular data to permit a much deeper understanding of the context of individuals’ driving behaviour in order to assess how likely they are to make a claim in the future. From their scores they can “differentiate loss ratio by over a factor of 10 from the best 10% of business to the worst 10%.” This consideration will be made more complex by the increasing array of data sources.