Fred Cripe is a senior advisor to PwC and a former executive vice president of Allstate Insurance Company.
Fred Cripe shares how the Internet of Things will shift the insurance business from loss compensation to loss control by helping customers achieve their goals.
Interview conducted by Vinod Baya
PwC: Fred, what is the history of telematics use in the auto insurance industry?
FC: The concept goes back a couple of decades, when Progressive filed patents related to the use of a tracking device in the automobile. It started with the thinking that tracking mileage was a better measure of risk exposure. Over a period of time, it became obvious to many in the industry that the accident rate is not proportional to mileage. The accident rate per mile actually drops off fairly steeply as annual mileage increases.
What the industry realized is that risk relates to how people drive. So the focus shifted more and more to using emerging technologies to track how people drove, where they drove, under what conditions, on what kind of roads, how congested the roads were, and so on.
PwC: Is such tracking becoming possible now?
FC: Cars already have a number of sensors that can provide information relating to how someone drives— information such as how fast they are going, how quickly they accelerate, how quickly they stop, what kind of turns they make, and so on. With advances in telematics, which is related to the broader trend called the Internet of Things, it’s becoming possible to get this information in batches or in real time at costs that are dropping.
From such information, and from knowing what the road conditions and speed limits are, you can in essence understand how well someone is driving. And that correlates with accidents. Obviously, the more dangerously someone drives, the more likely they are to be in an accident.
PwC: How is this information being used today and what is the long-term impact to the industry?
FC: Most carriers today focus simply on doing what insurers have historically done better, which is assess risk and charge people for it. Telematics can help determine whether someone is a better than average driver or a worse than average driver, so it is used as another input into pricing.
In the long run, insurance companies that really use telematics successfully will use it to change their customers’ exposure to loss. By charging customers for the kind of driving they do and communicating alternative behaviors that will cost less, these companies will give customers more control of their insurance costs. So if someone drives five miles on a dangerous road, that will cost the person more than driving five miles on a safe road. If someone engages in dangerous maneuvers such as hard braking, that will cost the person more than if they don’t. So in the long run, there is potential for far more sophisticated pricing, but that pricing enables consumers to have better visibility and control over their costs.
When I was at Allstate, we did some early trials with our own employees and agents to see what we could learn. We put sensors in cars that would measure how people drove, and the sensors also would give feedback whenever the driver made a risky move. Drivers would receive feedback by a little red light that would glow or by a sound that would say, “That was a risky move.”
In later tests of Allstate’s Drive Wise telematics product, they saw dramatic changes in driving behavior by employees. Whereas in the beginning only 25 percent of the testers scored in the ideal “safe zone,” over the course of the test that number increased to 75 percent.1 Providing real-time feedback promotes safer driving habits.
PwC: How will telematics change existing insurance company operations?
FC: In addition to loss control, I see telematics being used to transform our service, particularly in claims. When there’s an accident today, there’s enough information in the car from the accelerometer, airbag sensors, and other devices to indicate roughly the speed and direction of impact. We already know the particular make and model of the vehicle. With a set of predictive models, we can say with some confidence how much damage has been done, whether that car is going to be drivable, and the extent of possible injuries. And we can coordinate an appropriate response.
For instance, an insurance representative can call a customer’s mobile device in the car and check in: “We got the signal that you were in an accident. Is anybody injured?” We can provide necessary details: “Based on our assessment, your car is not safe to drive, and a tow truck is on the way.” If the customer has rental coverage, we can also send a replacement vehicle. “We’ll tow your car to this body shop and they’ll have an estimate by tomorrow.”
There is potential for much value-add. You can start to interview injured people immediately. You can get the police report in real time. The company maintains control of the damaged vehicle. Telematics will start to change how companies settle claims and serve their customers by making the overall experience more seamless.
PwC: What is the impact on the business models that are prevalent in the industry?
FC: By and large, today automobile insurance works as what I call the moving the money around business. Money goes in and then money comes out. Customers tend to think about whether they’re doing well or poorly in that relationship based on whether they’re getting more money out in claims relative to the money they’re putting in.
The business model change will move the industry from what I will call a reimbursement model to a prevention and loss control model. The basis of competition will switch from the ability to predict an individual customer’s expected loss to the ability to understand how a customer’s driving affects their expected loss and how changes in that driving—whether it’s where they drive, how well they drive, or when they drive—could change their expected loss.
PwC: Why is this ability possible today and not before?
FC: The business model I have described is only possible with a closed-loop system using sensors, real-time feedback, and predictive analysis of behavior data. We have had such a feedback loop for high-value assets. For example, in the property/casualty large commercial industry, part of the contract includes what is called loss control. If you had a factory, an insurance company would send experts to look at how your factory works and what kind of safety devices there were. They would also give you advice on how to how to run a safer factory, or have a safer store, or prevent theft, and so on. This advice is part of the service.
Such advice was always too expensive to provide for individuals. But with the Internet of Things, sensors, communications technology, and analytical power now make it feasible for insurers to offer loss control on an individualized scale for large numbers of customers. This capability extends to homeowners insurance as well. There are now devices that you can attach to pipes in your house that will let you know in minutes if a leak starts anywhere in the house. Also, you can plug a device into an electrical outlet, and it will let you know when a short circuit is likely in a particular circuit in the house.
PwC: Clearly there is much value-add potential to customers and the insurance industry. What are the challenges?
FC: First is a battle for customer acceptance. Only a very small part of the personal auto market is rated using telematics. Consumers really weren’t willing to have devices in their car that could track where they drove. There’s a major privacy issue about that. However, customer sentiments and expectations are changing, in part because of the smartphone. Customers expect more information and greater seamlessness in their experience, and telematics allows insurers to do that.
Today, about 1 percent of auto insurance customers use telematics. During the next 5 to 10 years, that percentage might grow 1 percent of the market a year. At some point it will skyrocket, and a critical mass from one-third to one-half of all auto insurance customers will be willing to purchase a telematics oriented product as opposed to the traditional insurance product.
PwC: As you look at the history and future of the insurance industry and the impact of the Internet of Things, what is the cosmic change here?
FC: The technology now enables companies to help customers achieve the goal that they’re buying the product for—as opposed to just selling it to them most cost-effectively, which is what businesses have done in the past.
In the case of the insurance industry, most customers don’t buy insurance to have a product. The insurance is to pay for the losses that occur when something bad happens to them. Their goal is to prevent a bad thing from happening to them in the first place.
Additionally, there is also alignment of customer and enterprise goals. Insurers can create more value in loss prevention for customers by saying, “I’m going to charge you 80 percent as much but reduce the likelihood you’ll be in an accident by 30 percent.” Insurers win because we pay out less relative to what we charge. The customers win because they save money and they really didn’t want the annoyance and the danger and the injury that comes with being in the accident in the first place.
1 Based on publicly reported data in the press release from Allstate Insurance, “Allstate announces crowdsourcing effort to test usage-based insurance product,” news release, July 25, 2012.