The insurance industry is a risky one; which is to say that successful insurance companies are ones that predict the risk of having to pay out on their policies. For car insurance, that means figuring out who’s a good driver and who speeds at 75 mph through a school zone with an open beer can in his hand while playing Candy Crush Saga on his iPhone. A company would want to charge hefty premiums to the second customer or maybe not even insure him at all, while it would encourage a good driver to join by offering discounts.
Since the dawn of the industry, companies have primarily used a driver’s record as a way of determining risk. How experienced are you? How many speeding tickets have you racked up? But now there is perhaps an even more accurate way to tell how someone is driving: the car itself.
“Internet of Things” technologies have profoundly changed many things, including the way that insurance is offered. Looking further into the future, self-driving car features could also alter insurance beyond recognition.
As cars (or bolt-on devices) become more and more loaded with sensors that measure speed, distance driven, braking, steering, signal use, and other points of data, insurers have begun to use this information to set premiums. Allstate, Progressive, and State Farm all offer devices that connect to vehicles’ on-board computers to monitor driving habits. Drivers can opt to voluntarily put these devices on their vehicles and collect a discount if they report good driving back to the insurer.
One of the pioneers of the “Internet of Things” revolution is Chris Brogan, a veteran of the telematics industry who was one of the first to use predictive analytics to improve safety. Brogan founded FleetRisk Advisors, Safety Intelligence Systems, and New York-based Assurenet, where he is currently CEO. These companies all focused on using data gathered from fleet vehicles, such as trucking companies or taxicabs, to see which vehicles were being driven in such a way that put them at risk for accidents. Assurenet also has plans for private auto insurers.
Brogan will speak at the TCF (Trenton Computer Festival) Saturday, March 19, at 3:40 p.m. at the College of New Jersey. Tickets are $12 in advance, $15 at the door. The festival, which runs from 9 a.m. to 5 p.m., will also feature a talk from Alfred Poor, publisher of Health Tech Insider, Ruth Wolfish of IEEE Client Services, Valerie Tucci, physical sciences and English librarian at the College of New Jersey, David Soll, Apple’s Dave Marra, and author Barry Burd, among many others. The festival will also feature technology demos and a tour of the Sarnoff Museum of Technology, For more information, visit www.tcf-nj.org.
Internet of Things (IOT) technology promises to make the insurance industry more effective at collecting data and predicting risk on all kinds of things it insures, as well as preventing losses in the first place. While auto insurance may be the most obvious place to use devices like this, insurers could also use environmental sensors in buildings to watch for smoke, toxin, and water damage. Many companies are also working on using “wearable” technology such as fitness trackers to help determine risk in the health insurance field.
Brogan has been at the forefront of the telematics business since the beginning. He grew up on Long Island, where his mother was a nurse and homemaker and his father was an engineer who eventually founded his own technology company. After serving in the Navy as a pilot of carrier-based planes, Brogan earned a bachelor’s in engineering from Manhattan College, attended Columbia’s executive MBA program, and followed in his father’s footsteps into the technology business.
After a stint with a startup educational software company, Brogan says he was hooked on technology entrepreneurship.
Brogan’s first company, Safety Intelligence Systems, founded in 2001, was based on an idea that proved to be ahead of its time. The company used Verisk’s “global safety data vault,” a massive database maintained by all of the country’s major insurance companies, combined with then-new “black box” data recorders for cars. The combination of these two technologies proved effective in a 2,000-unit trial run with the National Highway Safety Administration in Atlanta but was too expensive for broad commercial use. (GPS-capable tracking systems were $1,000 or more each at the time.)
But Brogan was just getting started in the safety business. His second company, Fleet Risk Advisors, focused on finding the drivers within trucking fleets who were most at risk for crashes. Its first client was Dupre Logistics, a Louisiana-based fuel hauling company with about 900 vehicles. CEO Reggie Dupre hired Brogan and his team to reduce crashes.
“Reggie Dupre was a big believer in safety,” Brogan said. “He had seen enough of his drivers killed while hauling fuel and enough people killed by his drivers hauling fuel. He told me he had gotten tired of going to funerals and had implemented some very cutting-edge safety concepts.” But the real improvement in safety came when Brogan’s company used the data from the Qualcomm trackers together with research from Circadian, a company that modeled human sleep patterns.
“They had a model of who would be tired and when based on your age, sex, weight, how many days you had worked, what your occupation was, whether you were a smoker or a drinker … all those factors determined your circadian rhythm,” Brogan said.
Brogan took every available bit of data from the black box trackers and from Dupre’s human resources department records and used it to create a spreadsheet of all 900 drivers in the company. He organized them by risk category, from the “green” drivers who were unlikely to have an accident, to a handful of drivers in the “black” who were fatigued and who were at imminent risk of crashing.
After using the model for a few months and refining it, Fleet Risk Advisors was able to predict 90 percent of all Dupre’s accidents. It turned out that predicting who was going to have an accident was more complicated than just seeing who had been at work for a long time. The most dangerous day was right after a vacation. “Vacation can be very exhausting, especially if you have kids,” Brogan said.
Dupre started using the data to take measures to prevent crashes. Drivers who moved into the “black” zone were given a two-day rest with no docked pay or other penalties. Drivers coming back from vacation were assigned to classes or non-driving duties for the first day back on the job. The changes were so effective that Fleet Risk Advisors ended up licensing the technology to Qualcomm under the name “Predictive Performance Service.”
In 2007 Brogan founded Assurenet to further develop Internet of Things-based safety analytics and apply it to new fields. “Our model became, ‘predict, prevent, record,’ Brogan said. Rather than create their own trackers, Assurenet has experimented with about 20 different telematics systems made by other companies. “There are so many guys out there building telematic systems,” Brogan said. “The ‘connected car’ space is extremely crowded.’”
Currently Assurenet is in the midst of a trial run of using sophisticated tracking devices, complete with video cameras, on a fleet of around 4,000 New York City livery cabs (essentially cars for hire.)
The company is also exploring getting into the Usage-Based Insurance (UBI) business with consumer auto insurance companies. Here in the Princeton-Route 1 corridor, other companies have already begun to incorporate UBI data into insurance policies.
New Jersey Manufacturers, headquartered on Sullivan Way in West Trenton, is in the midst of a pilot program that uses smartphones to track teen driving habits and give them feedback to become safer drivers. John Hardiman, NJM director of public affairs, said if the program is successful, the company will offer it to all of their subscribers.
Hardiman said the SafetySaves pilot program was offered to students at Bordentown high school and will soon be expanded to four more. Young drivers participating in the program can install an app, created by a company called D-Rive, that uses the phone’s GPS system to monitor speed, acceleration, braking, and cornering. The phone doesn’t offer instant feedback, but rather generates a report after the drive is over that can be viewed by the driver or his or her parents.
“It provides them with feedback and insights that they can take and use to determine where their weaknesses are and how they can become better drivers,” Hardiman said. Hardiman said the pilot program would help NJM determine if apps and smart devices could train drivers in safer behavior. “If the data indicates that in fact drivers get better and get safer, then it would stand to reason that there would be the potential for a premium decrease for drivers who used it.”
But not all insurance executives are convinced of the value of using the “big data” created by Internet of Things devices or smartphones. Eric Poe, CEO of CURE Auto Insurance based in Carnegie Center, pointed out one obvious problem with the devices, such as Progressive’s Snapshot dongle, that track driving habits: They are voluntary.
“Oh, you mean that Big Brother thing that Progressive has?” Poe said. “The reality is that it’s a marketing gimmick. Unless it’s mandated, you know that you’re going to get certain individuals who know their driving is impeccable using it. The people who are going to respond to that already know that they’re going to be driving within the parameters of the safety software.”
In other words, the devices will report back only what the insurance company already knew by virtue of the fact that the device was installed in the first place. On the other hand, habitual speeders and aggressive drivers would shy away from installing a dongle or a smartphone app.
Brogan agreed that the voluntary nature of UBI devices limited their value in a general pool of drivers, but that they would be very useful for high-risk groups such as teenagers, as NJM is trying.
Poe sees another downside to driver-monitoring dongles, which is fairness. Because the devices do not work in older cars that lack sophisticated on-board computers, they are unavailable to lower-income people who are more likely to drive obsolete cars. Fairness to lower-income customers may not be at the forefront of most business executives’ minds, but Poe is not a typical insurance executive and CURE is not a typical insurance company.
CURE was founded in 1990 by Poe’s mother, Lena Chang, an award-winning actuary, and his stepfather, James J. Sheeran, who was insurance commissioner of New Jersey between 1974 and 1982. Poe went to law school at Seton Hall and went back to school to study a subject he hated: accounting. (Poe said his birth father, who was Chinese and an ardent follower of Confucian philosophy, always told Poe to work hardest at the things he loathed.)
CURE operates differently from other insurance companies in that it only uses driving records to set its premiums, rather than education, credit score, or other measures that Poe says most other companies use as proxies for income. (Wealthy customers are more profitable for insurance companies because they are likely to buy other insurance products and insure more expensive things.) Poe has long lobbied to change the law to prevent credit ratings or income proxies from being used to set premiums, arguing that it’s unfair to make lower-income people pay higher premiums.
Poe is suspicious of Internet of Things devices as being less objective than they first appear, due to the dynamics of who can afford compatible cars or phones. However, he said, he would approve of them either if they were mandatory, such as if a law were to be passed requiring them, or if a fleet installed them in all of its vehicles.
“If you mandated a customer use it at renewal, like okay, I’m going to increase the rates for these seven drivers out of 500 because they’re driving terribly, then it makes sense,” he said. “But those individuals are never going to say ‘I want to be monitored.’”
Not all business leaders are as skeptical as Poe, and many have embraced Internet-connected monitoring devices. Business Insider Intelligence estimates that by 2020, 50 million drivers will use UBI (usage-based insurance) devices.
In some ways the addition of plug-in tracking devices to cars is redundant because cars themselves are equipped with dozens or more computers and trackers right out of the factory. Every system on modern cars is computer controlled, from the air-to-fuel mixture in the engine to the stability control system to the radio. Event Data Recorders, common in cars since the early 2000s, continuously record information from all vehicle systems and save the last few minutes of it in the event of a crash, functioning much like an airplane’s “black box,” to help reconstruct accidents.
The proliferation of this kind of driving data has sparked debate over who owns the data: the driver or the car manufacturer? A bill introduced in the Senate last year, the Driver Privacy Act of 2015, would settle the issue legally by declaring that black box crash data would be the property of the owner or lessee of the car and could only be accessed by third parties under certain circumstances such as a court order.
“Connected Car” systems like GM’s OnStar allow drivers to control their vehicles remotely through the Internet, and the connectivity has reached a point where security experts are worried that cars are vulnerable to hacking. In March, 2015, a pair of hackers demonstrated they could breach the security of a Jeep Cherokee and shut down its engine while on the highway. The Jeeps were recalled as a result.
Looking farther into the future, self-driving car technology could overturn the entire idea of insuring a “driver.” What’s to insure if the operator of a car is merely telling the car where to go and letting the vehicle make all the decisions itself?
Carmakers and technology companies are investing heavily in self-driving car technology. Electric car manufacturer Tesla already offers an “autopilot” function on its luxury sedans that is like an advanced cruise control, allowing the cars to cruise down a highway without the driver steering or using the controls at all.
Google has been testing a fleet of small self-driving cars. In February, one of its vehicles passed a dubious milestone when it was found at fault in a collision, running into a bus at two miles per hour. Ten American automakers have volunteered to make automatic braking crash avoidance systems standard in all of their cars.
The question of how to insure cars with various levels of self-driving capability is one that all companies that offer auto insurance will soon have to face if the industry continues to develop at its rapid clip.
In the short run, Poe says that self-driving cars will make the job of an insurance company much easier. Because self-driving cars would come equipped with cameras and data recorders that would show the exact circumstances of an accident, determining liability in a car crash would be a snap. The company would know either that their car wasn’t at fault, or if it was, if they should settle a case early instead of wasting time defending it.
“It’s very hard to find credible evidence aside from impact points when you’re involved in a car accident,” Poe said. “If you’re going to give me the ability to track movements and to look at actual video of an accident, I would be able to better ascertain my liability.”
Brogan said that determining liability is one of the primary uses of the tracking systems used in Assurenet’s taxicab program, especially in the contentious “no-fault” legal environment of New York City, where insurance fraud is an ever present risk. Cameras showing exactly what happened in an accident have proven invaluable for drivers and the cab company because they provide solid evidence of who was at fault in any accident. The company then knows whether to settle a case or defend it.
In the long run, if and when robot drivers rule the roads, Poe anticipates that motor travel will be so safe that insurance premiums will drop dramatically, or even reach the point when car insurance is no longer necessary at all.
Poe is skeptical that humans will be willing to hand over the wheel any time soon. Self-driving features are currently only available on the most expensive cars, and even if they became commonplace, it would take decades for the fleet to turn over. “I do not think most people would give up the individual freedom of operating a motor vehicle,” Poe said.
Neither Poe nor Hardiman foresee a rapid switch over to self-driving technology, but rather the slow addition of self-driving features to new cars. Hardiman said the insurance industry is anticipating the rate of accidents to gradually go down as these safety features, such as automatic braking, are introduced, and that premiums will go down accordingly.
However, Poe said there have been unintended consequences of using increasingly sophisticated safety features in cars. Airbags, for example, have made cars safer for their occupants and reduced the number of injuries and deaths in auto accidents. But with new cars coming equipped with a dozen or more airbags, the property damage incurred by accidents has skyrocketed. Poe said each airbag that goes off in a collision can cost thousands of dollars to repair. A fender bender could end up easily totaling a new car.
Likewise, the addition of fancy cameras and sensors to vehicles could further drive up the cost of repairs and increase the cost of collisions when they do happen. Poe said manufacturers have an incentive to make their cars more expensive to repair since dealerships make far more on servicing vehicles than by selling them.