When I was growing up, my dad kept a logbook in the glove compartment of his car to record how often he filled the tank, how much gas he used, and how much it cost him. The glovebox also held a little envelope filled with each month’s gas receipts, which he tucked into the back cover of the log. I did the same for a while, as well — it made me feel like a responsible car owner.
“I did it, too — it was a very common habit for a long time, especially for men,” says Ljuba Miljkovic, chief product officer of tech startup Automatic, which sells a gadget that tracks and scores driving habits with an eye toward improving fuel efficiency. “And a surprising number of people still do it because they learned that was the ‘right’ way to own a car. But when we asked people who kept these logbooks whether they ever actually looked at them, the answer was largely no.”
Miljkovic likens this analog version of collecting data to no real end to many of the first wave of apps designed for the so-called “quantified self” movement. These apps gather data about various personal habits and the state of your health in order to prod you into or away from activities.
Automatic, on the other hand, aims to use data points about driving behavior — how often and how hard you brake or accelerate, for example — to help people drive more safely and conserve fuel. There’s no logbook, whether analog or in the form of an app in which you tap away at measurements. Automatic does this all, well, automatically.
By replacing a manual task with gizmos that not only gather and analyze data points but also suggest small shifts in behavior based on that data, Automatic is following an ever more popular school of thought about behavioral changes in which technology provides the key.
Ambiguous by design
Two years ago, Miljkovic attended a lecture given by experimental psychologist B.J. Fogg, who heads up Stanford University’s Persuasive Technology Lab. Fogg coined the term “captology” to describe technology designed with the purpose of changing people’s attitudes or behaviors.
In the lecture Miljkovic attended, Fogg laid out his behavior model for designing persuasive technologies, which posits that in order for a technology to influence behavior, it must have a convergence of three elements: motivation, ability, and trigger.
In an interview, Fogg points to Instagram (which happened to be started by one of his students) as a great example. “It’s a brilliant piece of behavior design, and there are a few reasons that it’s so successful: first, it helps people do what they already want to do — show the coolest version of themselves to the world; second, it’s really easy to use; and third, it lets them put a filter on it that enables them to show how artistic they are and to add perspective to an image,” he says. “It’s making it really easy to do something people already want to do. Everything that’s winning in behavior design right now fits that pattern.”
It was with this model in mind that Miljkovic came up with the idea to build real-time audio feedback into Automatic. The app connects via low-power Bluetooth 4.0 to the Automatic Link, a small device that plugs into and is powered by the OBD-II data port that’s found in every American car sold since 1996. The Link gathers real-time data; the app uploads those metrics to Automatic’s cloud for processing (and backup) and displays the analyzed results.
The Link sounds a subtle alarm when the app decides you’ve done something that could be described as “aggressive driving.” Different tones signal which avoidable behavior just occurred. “We really weren’t sure what people’s motivation would be to change their driving habits,” Miljkovic says. “We knew they had the ability, but we also knew that there are various ‘costs’ associated with any change — maybe someone is worried about going too slowly, for example.”
Miljkovic adds that because everyone thinks of themselves as a good driver, it was important to move from offering feedback about good or bad actions toward just making drivers aware that, for example, the stop they just made was more aggressive than it needed to be, and that braking more smoothly would have saved gas.
“Our real-time audio feedback is a trigger — not necessarily for immediate behavior change, but to form a memory and inform what you do in the future,” Miljkovic explains. “When we tested real-time feedback, we found that behaviors people had been doing forever and not really thinking about became something they actually thought about changing.”
The approach appears to have hit a nerve. Automatic’s first product, a combination of the Link and the app that retails for a one-time fee of $99, was available for pre-order in March 2013 and outsold its first two months’ estimated production supply in the first week.
The product officially rolled out in June 2013, and by November Automatic was announcing the inclusion of its product in both offline and online Apple stores, a rare feat for any non-Apple product, much less a brand new one.1
Change isn’t always good
Automatic’s founders, Thejo Kote and Jerry Jariyasunant, met at the University of California, Berkeley, while researching the costs and trade-offs of different transportation modes and how one might go about nudging people toward one mode or another. Because so many people still prefer driving, yet so few people understand how much their driving behavior affects their fuel usage, Kote and Jariyasunant decided to focus on driving behavior. Automatic was born.
The sort of deep knowledge that Kote and Jariyasunant had of transportation forms a key differentiator between Automatic and most other persuasive technologies that aim to deliver an environmental benefit, according to Fogg. “The people who understand which behaviors are the right ones to target are not the same people creating the innovations, and there’s not a good bridge or exchange between those two camps,” he says. “The same thing is happening in health, which is why you keep seeing these articles about how this or that health app is a failure.”
That leads to designers who are often just guessing at which behaviors to target, and products that wind up shifting either the wrong behavior or a behavior that just doesn’t have much impact. Recycling is a prime example. It’s largely seen as one of the more successful attempts at changing behavior. People have been retrained on a mass scale to perform a sorting operation, rather than throw everything in the trash. However, recycling shifts the symptomatic behavior of handing off used goods or remains, rather than addressing the root behavioral problem: the over-consumption of disposable goods.
Fogg also points to recent interest around tech products that encourage people to reduce their shower times as an example of the problem with most technologies geared toward changing human behavior in a way that benefits the environment. “Shorter showers may deliver real environmental benefit, I don’t know, but the problem here is that the designers don’t know either,” Fogg says. “Some computer scientists came up with this idea of, ‘Hey, let’s get people to take shorter showers,’ and they didn’t talk to any environmental experts, but their colleagues said, ‘Yeah, that sounds great,’ and now there’s a whole program around that, and full industry research projects,” he says.
When priorities compete, the environment loses
Part of Automatic’s success is its ability to address multiple priorities without pitting one against the other. “The behaviors we measure for fuel efficiency — speeding, rapid braking, and so forth — all correlate with unsafe driving as well,” says Miljkovic. “We can meet people where they are and help them achieve the goals they set themselves, whether it’s efficiency for environmental reasons, or because they want to save money, or it’s really just safer driving.”
That turns out to be crucial to the success of any technology designed to shift human behaviors related to the environment, according to Fogg. “In the environmental realm so far, it’s mostly been technology that forces you to do something — rather than technology that encourages you to choose a better behavior — that has been effective,” he says.
Eco-friendly behaviors tend not to have any sort of immediate reward. “That’s why the Prius and certain kinds of hemp this, eco shirt that, do well,” Fogg says. “There is actually an immediate reward there: people get to look good to their friends. If you take away the social approval and the social judgment, those products aren’t as successful.”
To combat the lack of immediate gratification connected to many environmentally responsible actions, Fogg says designers need to relate the desired eco-friendly behavior to other, more immediate human priorities. “People don’t hate the environment; they care about it,” he says. “But in the moment, they often care about other things more, and that’s how they decide. It’s a matter of competing motivators, competing priorities.”
Speaking from his car during a long, congested commute from Sonoma to Palo Alto, Fogg cites a proposal for a train along the same route as an example. “I’m driving on a highway right now where in two years they’ll have a train,” he says. “I’m going about 15 miles an hour in the carpool lane. If a train existed, people would ride it to save time, not to save the environment, but both goals would be accomplished.”
Small shifts lead to broad strokes
While Automatic is helping one driver at a time save fuel and avoid accidents, a broader technology movement aimed at changing behaviors around personal transportation is hoping to have a larger environmental impact. Called Mobility-on-Demand (MoD), this combination of driverless cars, software that analyzes transportation supply-and-demand, and “smart” electric vehicle-charging infrastructure aims to eliminate the need for personal car ownership in urban areas. And although MoD would dramatically reduce greenhouse gas emissions, its proponents tend to focus on more immediate, tangible benefits.
“A well-designed Mobility-on-Demand system should provide more convenient, flexible, and less expensive alternatives to private car ownership,” says Ryan Chin, the managing director of the City Science Initiative at the Massachusetts Institute of Technology (MIT) and a longtime leader in the MoD space. “I don’t say ‘more sustainable,’ because at the present time, only the highly educated typically value that quality. The rest of society doesn’t, primarily because we don’t currently pay for the environmental damage caused by using fossil fuel-based automotive transportation.”
Until that changes, Chin and company continue to focus on making MoD convenient and economically attractive via new vehicles, smart charging systems, and trip planners. On the vehicle front, MIT — along with several others, including Google, BMW, and Mercedes — is working on several versions of autonomous cars, which Chin describes as driver-less taxis that would solve the so-called “last-mile problem” of public transit by providing transportation between people’s homes or workplaces and public transit hubs.
MoD is sometimes confused with personal rapid transit (PRT) — an idea that has been proposed and then abandoned a few times over the past few decades — which relies on individual public transit pods that would transport people throughout a city using extensive monorail systems. Although PRT has been successfully deployed at some airports, it has fallen flat as a solution for broader transportation issues in no small part because the idea of crisscrossing monorail lines covering urban skylines has proven to be unappealing to residents.
With MoD there’s no major new infrastructure required and no meaningful shift to the ways people are used to getting around. Electric MoD cars would be governed by trip-planning software that ensures that cars are always close enough that no one has to wait more than a couple of minutes for one, that the cars can park themselves in central locations when not in use, and that idling (and thus energy waste) is kept to a minimum.
Until MoD is available, however, a handful of interim solutions intend to change people’s perception of on-demand personal transportation. Bike-sharing is finally taking root in American cities, with major rollouts in New York City and San Francisco in 2013.2 German company car2go has made its reservation-free car-sharing service widely available in the U.S. Unlike Zipcar and similar efforts, the company lets its members pick up a car wherever they find one and release it in nearly any paid or free parking spot within a metropolitan area. Members don’t pay for the time they aren’t using a car, and they don’t have to return it to its point of origin.
In Las Vegas, Project 100, part of Zappos CEO Tony Hsieh’s Downtown Project, will launch in early 2014 and provide bike-sharing, car-sharing, public transit, and an Uber-style towncar service (using Tesla Model S sedans) to members, who will pay a flat rate of $400 to have the same go-wherever-whenever access to transportation that personal car ownership affords.
“The ultimate idea is to have a fleet of self-driving electric taxis roaming around moving people in cities,” Chin says. “When we get to fully autonomous vehicles and systems, we’ll be equipped to reap the full benefit of them. With an autonomous vehicle, you could even drink and drive!”
What was that about making it easy for people to do what they already want to do? Safe and legal drunk driving may be just the trigger needed to get people to give up their cars.
Top photo courtesy BMW; bottom, courtesy Automatic.
The Automatic Link joins a number of other, wonkier OBD-II–compatible devices that are used by people who like to quantify their driving or tweak their automobiles. They will soon be joined by a competitor that sends data via a cellular connection directly to cloud services; it has reached funding on Kickstarter, and it was developed by the former CIO of the state of Utah, Phil Windley. ↩
An independent journalist based in Oakland, California, Amy Westervelt writes about health, technology, and the environment for a variety of publications, most recently The Wall Street Journal, Fast Company and Forbes.