UBI should really stand for User-Based Insurance, not Usage-Based Insurance. This small change in definition would immediately expand the very narrow view of what UBI should mean for car insurers. UBI as defined today should actually be called VBI – Vehicle Based Insurance. Most insurers only take into account how people handle their vehicle. They ignore the entire context of the user or the driver. Real UBI (ie UsER Based Insurance) solutions put the user at the core, not the usage of a vehicle (aka UsAGE Based Insurance). They put the user at the center by looking at the entire context of a user’s behavior. This includes in-car focus, origins and destinations (O&D’s) and lifestyle.
Driver’s Activity Focus
What else is the user doing in the car when they drive? Are they using their mobile phone? Mobile phone usage has become the number one cause of car accidents in the US. Real UBI solutions have to track drivers’ cellphone related distraction (or focus). Only a few SDK based UBI solutions track this accurately.
Origins and Destinations (O&Ds)
Where are drivers coming from and where are they going. Someone who drives home from a bar every night probably has a higher risk profile than someone who drives home straight from work. Knowing the before and after trip activities add highly relevant contextual insights.
Understanding the entire context of a user is not only important for making more accurate predictions about risk profiles. It also allows insurers to coach their customers to become better drivers. If you want to coach drivers you don’t just need to know how they drive, you need to know why they drive the way they do. For coaching, context is almost more important than the actual driving behavior.
So how can insurers make the transition from VBI to real UBI? There are 5 things insurers need to take into account when designing their next generation UBI solutions.
Telematics-based solutions will never get us to real UBI. They will always be confined to the vehicle. The only path to real UBI is SDK based solutions that can track the broader context of the user. The main argument against SDK based solutions has traditionally been accuracy. It is true that a few years ago Telematics solutions were more accurate. But a lot of progress has been made recently. Artificial intelligence and increasingly deep learning algorithms have become incredibly powerful in filling in the missing data holes SDK based solutions used to struggle with. A large global ride sharing company recently tested Sentiance’s SDK based solution by comparing it side by side to an IMU box (a very sophisticated device used to calibrate Formula 1 cars). The test showed a 97% accuracy of the Sentiance SDK based solution. Not every SDK based solution will be as accurate. In fact, there are still very large differences in accuracy between different vendors. But an insurer can easily set up a similar side-by-side test to ensure the solutions meet the required accuracy thresholds.
In order to protect the privacy of their customers, insurers should be explicit about what they track and should never track anything beyond low level, anonymous signals produced by the sensors in the mobile device (typically location, time, speed, acceleration and gyroscope data). The best solutions will only use this very limited set of parameters to infer the broader context about the user. This requires very advanced data science capabilities only a few solutions provide. Doing more with less data is the game here.
UBI solution providers will never know exactly how an insurer’s actuarial algorithms work. So they need to be open to work with the actuaries on the insurer’s side to tailor the outputs from the algorithms to the insurer’s needs. Some solution providers only hand over one driver score. Others open up the entire dataset so the actuaries can test which variables are most predictive for their purposes.
VBI is car-centric while UBI is driver-centric meaning, profiling individual drivers vs. single cars. This way we can distinguish multiple drivers sharing the same car, but also individual drivers using multiple cars.
One of the main advantages of SDK based solutions has always been the ease and speed with which they can be deployed. Most SDK based solutions still require a small Bluetooth device that tells the app the driver is in the car. But the cost of this device is negligible compared to telematics solutions. In the not so distant future, we will probably no longer need these Bluetooth devices. The algorithms will be able to detect a driver’s DNA based on their behavior, which will allow them to determine if the user is driving the vehicle or if she is a passenger. This will increase consumer adoption even further. The installation of hardware in cars is often a mental barrier for consumers.
SDK solutions are not only more easily deployed, they are also cheaper as they don’t require any hardware installation. But even among SDK based solutions there are high variances in cost. This is because vendors continuously push the science on how to reduce computing costs, the main driver of the overall cost of the solutions. This means that some SDK based vendors have been able to reduce costs by 50% while maintaining the same levels of accuracy. Car insurers should therefore do their due diligence when selecting a vendor.
Usage-Based Insurance (UBI) is really getting to a point of critical mass. Adoption of UBI products is rising fast, especially among new customers. At a recent conference in Chicago Nationwide said that in June 2016 36% of their new customer enrollments included some element of UBI. That number was up from 11% in 2013. As this market reaches maturity insurers need to make sure they move to the next generation solutions that put the user at the core. It’s time to put the right U in UBI.
At Sentiance, we developed a sophisticated platform that allows us to leverage smartphone sensors to overcome these disadvantages. Smartphones contain the exact same sensors as costly black boxes, namely accelerometer, gyroscope, magnetometer and GPS sensors. These sensors are used by the phone to automatically flip the screen when the phone’s orientation changes, to augment mapping and routing applications, and to improve the gaming experience. Read more about how we model driving behavior using smartphone sensor data.