A computer network designer takes on commute-hour gridlock in three countries.
Traffic in Bangalore (Bloomberg/Getty Images photo by Namas Bhojani)
Stanford computer scientist Balaji Prabhakar was on his way to a routine business meeting in Bangalore, India, when his career took an interesting detour. “It was the winter of 2007,” he recalls. “I was being driven to Infosys, which has an IT services campus on the outskirts of the city, when suddenly I found myself in the mother of all traffic jams.” Hemmed in by honking cars, trucks, bicycles, and scooters, Prabhakar typed out an apologetic text to the person he was supposed to meet. Then he posed what he thought was a rhetorical question: “Doesn’t anybody care about this traffic?!”
In fact, Infosys executives had been thinking about traffic gridlock a lot. Like many booming Indian IT companies, its Bangalore campus is located in a special Indian economic zone called Electronics City, about 15 kilometers south of the city center. Every morning, more than half of its 20,000 employees have a nightmare commute indeed. So when Prabhakar finally arrived at his destination, the execs asked the computer science/electrical engineering professor what he would do to make things better.
Prabhakar, who grew up in Bangalore, came up with a novel solution: a high-tech “nudge engine” aimed at encouraging Infosys employees to come in early and beat the morning rush. The idea worked so well, it is now being used to tackle similar problems in cities around the world.
How does it work? Nudge engines are programs that use mobile, cloud, and social networking technologies to sense individual behaviors — the number of times employees swipe their identification badges at work, for example — and then “nudge” them to change those behaviors through the use of friendly competitions and incentives, such as lottery prizes.
The idea comes from a basic tenet of computer science: that shifting just 10% of the peak load in a congested network can lead to a dramatic reduction in congestion measures, such as waiting times. It also builds on the work of UC Berkeley economist John Morgan, who has suggested that lotteries can be an equitable and effective means of advancing the public good without the downsides of fines or differential pricing schemes.
In Bangalore, for example, employees who signed up for the Infosys Stanford Traffic Project (INSTANT, for short) were monitored to see when they arrived at work each day. Rather than earning token amounts for swiping in early (which would have been a ho-hum scheme), early birds were given credits toward a highly anticipated Saturday morning online lottery, with cash prizes ranging from the Indian equivalent of $10 to more than $200. The more early bird trips individual commuters made, the higher the chances they would win, and the higher the potential rewards.
The approach worked. By the end of the 6-month experiment, the number of Infosys commuters arriving before 8:30 in the morning had doubled, and the average bus rider’s commute fell from 71 minutes to under an hour, allowing Infosys to shift 60 of its shuttle buses to earlier end-of-workday departure times.
Stanford computer scientist Balaji Prabhakar came up with a high-tech solution to help employees in India beat the morning rush on their commute to work. (Associated Press photo by Rafiq Maqbool)
The program also turned out to be great for morale. “INSTANT brought people together,” says Rama N.S., former head of Infosys’s Bangalore Development Centre and now CEO of the Electronics City Industry Association. “It was fun, it reduced congestion, and there was money to win. The algorithm was robust, and people said it was very fair.”
Prabhakar and his Stanford graduate students had thought that the Bangalore experiment would be a fun, temporary diversion from their regular work in Stanford’s departments of electrical engineering and computer science. A Stanford faculty member since 1999, his teaching and research have focused on the design, analysis, and implementation of data networks, both wired and wireless. But he saw the intriguing similarities between computer and phone networks, which have limited bandwidth, and real world networks, such as transportation systems and electrical grids. “When a network is successful like the internet, and it grows to billions of users, it’s very difficult to change it from the top down,” he notes. Likewise, “you can’t build new road systems overnight. But you can try small changes. That’s the beauty of incentive programs. You don’t have to start changing everything on day one. You can start small, and scale out.”
As the project gained more attention, including front page coverage in the Times of India, it attracted the interest of curious transportation managers in other parts of the world. One of the callers, from Singapore’s Land Transport Authority, wanted to encourage train riders to shift their commute schedules away from overcrowded peak times through the use of incentives. So after meeting with the authorities there, Prabhakar and his students devised a new nudge engine called Insinc, giving participants points for every kilometer traveled on the system, and triple points for off-peak morning trips.
The interface looks like a combination of the popular Chutes and Ladders board game and a program for frequent flyers. (Committed off-peak train riders can work all the way up from bronze to platinum status.) When a Singapore commuter spins the virtual wheel, a token car moves onscreen. Players who land on lucky squares can win prizes of up to $200.
Prabhakar and his students launched a similar nudge engine aimed at reducing daytime automobile congestion and parking problems on the Stanford University campus in 2012. Funded by a grant from the U.S. Department of Transportation, the Capri program lets regular Stanford parking permit holders earn incentive points whenever they bike or walk across campus borders, or drive in and out at designated off-peak hours. Solar-powered sensors, similar to electronic toll booths, were installed at various campus entry points to keep track of the comings and goings.
Currently, about half of the 8,000 Stanford employees and eligible students are participating in the program for rewards ranging from $5 cash prizes to highly coveted sports tickets. “I hear that it’s common for cars to be lined up outside of the sensors at five minutes to six in the evening, just waiting for the time to change,” Prabhakar says, grinning. So far, he says, about 15% of the trips taken by participants have shifted away from rush hour.
Another of Prabhakar’s nudge engines didn’t have anything to do with traffic. Instead, it aimed to boost the physical fitness levels of employees at Accenture USA, the management consulting, technology services, and outsourcing company. To participate, employees uploaded pedometer readings into the “Steptacular” database, which awarded credits based on steps taken. Employees then redeemed the credits for random rewards through an online user interface.
A particularly effective feature of the Steptacular program was its use of social networks to nudge employee behavior. Looking at the website, participants could see how many steps their colleagues had taken, and how many steps their entire office had taken compared to others in their company. “The graphic showed a rocket going from the earth to the moon,” Prabhakar explains, “and the distance it covered was equal to the total miles walked by all the employees in the program, with contrails showing the distance covered by regional offices.” Coding the playful program was an enjoyable change of pace for the professor and his graduate students. “We were here in the office all the time!” he recalls, laughing.
In the future, Prabhakar says, nudge engines could be used for a variety of purposes: to encourage people to install computer security patches, to recycle waste, to save electricity, or to conserve water by taking shorter showers. “You’d have to sensor it, maybe by wearing a small tag when you enter the shower, or by sensing the footprint on a bath mat,” he muses. In any case, he adds, “The idea is to encourage desirable behavior shifts — and make it fun.”
Prabhakar described the Bangalore project, as well as related experiments in Singapore and the Bay Area, during a 2014 winter quarter course called Designing Large-scale Nudge Engines, at Stanford Graduate School of Business.
For media inquiries, visit the Newsroom.
Explore More
How to Chat with Bots: The Secrets to Getting the Information You Need from AI
You’re in Charge: How to Use AI as a Powerful Decision-Making Tool
