Congestion is a common problem in services. A large number of customers put demands on the system all at the same time and delays ensue. A few weeks ago we posted about GymFlow, an app that tries to address congestion at health clubs by providing better information. GymFlow doesn’t tell you can’t go to the gym at 5:30. It just points out that the gym is going to be a whole lot less crowded if you got 3:30.
Now the Wall Street Journal has an article on a different way to ease congestion by relying on games and lotteries (Gaming the System to Beat Rush-Hour Traffic, Aug 1). It reports on the work of Balaji Prabhakar, a Stanford Computer Science professor, who has tested out various systems to get commuters to tweak their travel habits. The article’s author discusses his approach here:[audio http://podcast.mktw.net/wsj/audio/20130801/pod-wsjwnwesselcapital/pod-wsjwnwesselcapital.mp3]
Here is a summary of one of Prabhakar’s at his place of employ.
His team recently brought the technique home with a federally funded experiment to help Stanford keep its promise to Santa Clara County to alleviate rush-hour traffic. The 3,900 participants—a significant share of the relevant pool of 8,000 parking-permit holders—installed devices on their cars (soon to be replaced with a smartphone app) and got points for arriving and leaving an hour before or after the rush hour.
The popularity of the Chutes & Ladders-like game stunned Stanford’s director of parking and transportation, Brodie Hamilton. He doubted people would take the time to spin the electronic dice to play it, and insisted that Mr. Prabhakar include an auto-play feature. But, Mr. Hamilton says, “I have people on my staff who play it regularly. People are really into it. Balaji was right!”
About 15% of the trips taken by participants have shifted away from rush hour. Students tend to come and leave later; staff tend to come and leave earlier. Smartphones make all this easier to implement: A new mobile app tracks bikers and walkers and gives them points, too.
Those who commuted off-peak got points to play in the on-line game with a chance to win cash. We are not exactly talking a year’s tuition here. The program’s website touts “random cash rewards from $2 to $50.”
So one way of looking at these programs is to say that they are peak load pricing schemes. If you are willing to travel at some non-peak time, you get a discount. In a case like Stanford in which there is no explicit toll, that discount becomes an outright reward. The twist is that the rewards are random. On average, everyone gets a positive reward. In reality, a lot of people get nothing while one person wins $50.
One might argue that these program are a way to take advantage of people. They are inducing commuters to incur a cost to adjust their schedules for little or no return. However, they are not forcing anyone to participate and these game schemes are less coercive than raising the peak price for everyone. For example, a public transit system might impose a rush hour premium in order to shift some demand to earlier times. But how big does that premium have to be to sway the behavior of a lot of people and what happens to those who cannot adjust their schedule? The article reports that Seattle’s transit system impose a 75¢ premium for travel between 6:00 and 9:00 in the morning. That amounts to less than $200 per year. I have to think that all but the most price sensitive/budget conscious just ignore the fee. But even the cost conscious might be forced to suck it up and pay. If a parent can’t leave the house until the youngest is off to school, leaving before 6:00 is not an option. Implementing peak load pricing is then limited by the political reality that a premium high enough to really sway behavior may be impossible to implement. This gamification approach can scout the negative impact of a high peak price while (apparently) still roping in enough commuters to make a difference.
The one question I have on this is how long the effect lasts. By design a lot of people ain’t going to win much. So what is the incentive to stick with the game over time? Making it easy to participate (e.g., Stanford’s auto play feature) will help, but I have to think that this is out-of-sight out-of-mind and all but a die-hard few stop carrying about the program.