A traffic jam on the web typically begins and ends in less than a millisecond. But the Internet faces many of the same roadblocks as our roadways. The biggest: that we tend to commute and browse en masse at certain points in the day creating moments of intense strain on the infrastructure after long periods of underuse.
The big internet companies have invested heavily in architecting systems that ensure a moment of intense traffic doesn’t cause a pile up. Now, a handful of ex-engineers from Google and a specialist from Stanford want to take their learning from playing traffic cop on the web to solve the congestion problems plaguing some of the world’s business cities.
The company, Urban Engines, uses a combination of fare data and economic theory to help cities such as Bangalore, Aao Paulo and Washington D.C. reduce congestion on trains and buses during rush hour. Municipalities feed fare data into the system, and the company uses machine learning to create visualization and insights that help operators identify where and when to add trains, buses or other capacity to their networks.
“We realized that what folks were doing to alleviate traffic in the real world was similar to the early days of managing infrastructure on the web,” said Karen Roter Davis, an Google veteran who now serves as general manager at Urban Engines. “What we asked was whether we could apply the same theory in the digital world to the real world.”
The answer was yes with a caveat: humans, unlike packets of information, have free will. A engineer can route data through which ever server she chooses, but convincing a commuter to leave ten minutes later to avoid rush hour requires finesse.
So the company developed a system through its website that offers commuters a small incentive to leave a few minutes later or earlier depending on their commute. By analyzing the traffic patterns gleaned from the fare data, the company can identify the routes which contribute most to the spikes in congestion. Then, the company offers small reward incentivizing register users to switch up their commute, effectively reducing the strain during peak times.
“One of the first things we realized was that offering someone ten cents will not cause people to shift behavior,” said Davis. “But if we pool those dimes together into a micro-raffle, and we say if you shift you could earn points, its far more effective. People are much more likely to make a small shift for a potentially large reward, than a definite small one.”
The early tests suggest a portion of commuters may be willing to change their habits. In a small pilot in Singapore, run a year ago before Davis joined the team, the company was able to reduce the variation in congestion between peak and offpeak, a measure of efficiency, by 7-13% in less than 6 months — all by better managing the relationship between supply, the buses and trains, and demand, the people who cause the congestion.
Today, the company works exclusively with public transportation systems, but Davis suggested the technology could eventually apply to a variety of business who rely on moving people through places. Few business operate at the scale of public transportation systems, but retailers, food service and place-based business will increasingly collect meaningful amounts of movement data through mobile applications.
The urban congestion problem is a behemoth, and the startup may never move into retail or other categories. But the the project is another example of the rapid exporting of web technologies, and expectations, to the real-world markets.
More and more, we expect our interactions in the physical world to match the immediacy and seamlessness of the digital experience. The long lines at a coffee shop or overcrowded trains, which we once presumed an inescapable reality of daily life, seem solvable and avoidable. Waiting is something we already despise, and technology is teaching us that it is unnecessary.
As consumers gain more choice and transparency in their local markets, the margins by which a consumer picks one seller over the other will lessen and user experience design will become more than a software strategy. At some point soon, the batch of offline operations technologies — everything from in-store analytics to booking systems — will transition from a cost reduction strategy to consumer experience mandate.
Steven Jacobs is Street Fight’s deputy editor.