PLAN 75: Article
A Winning Approach to Social Networking
A team from the Media Lab’s Human Dynamics Research Group recently scored a $40K prize in a social networking contest sponsored by the Defense Advanced Research Projects Agency. The tools they developed to win the contest could turn out to be important in large-scale collective problem solving such as finding a missing child, spreading information about a product safety recall or mobilizing rescue efforts during natural disasters such as earthquakes or tsunamis.
Staged to celebrate the 40th anniversary of ARPANet, the precursor to the Internet, the contest challenged participants to use social media to locate ten red weather balloons placed at undisclosed locations across the United States. From a field of more than 4000 teams, the MIT winners included graduate students Anmol Madan, Wei Pan and Galen Pickard along with postdoctoral fellows Manuel Cebrian and Riley Crane.
The crux of the team’s approach was the incentive structure it designed, a way of aligning the interests of participating individuals with the interests of the group. Whoever provided correct coordinates for one of the balloons got $2K; whoever invited that person to join the network got $1K; whoever invited that person got $500, and so on. No matter how long the chains got, the total payment for all ten balloons never quite reached $40K, and whatever was left over went to charity. In addition to the monetary rewards, the system also allowed participants to see their direct impact on the network.
The design of the scheme meant the researchers didn’t get any of the prize money but the team’s real motivation was to explore how to mobilize the vast resources of the human network and exploit the opportunities that come with living in such a connected world. According to team member Riley Crane, ‘we believe we have developed a set of tools that can find the proverbial needle in a haystack’.
The team assembled its strategy in only four days. On Wednesday, they started discussing potential approaches to the problem, by Thursday they had built a demo of the website they used to aggregate data and on Thursday evening the site went live. Within two days, 5000 people formally joined the team’s network and by 6:52PM on Saturday – only 8 hours and 52 minutes after the contest began – the team had located all ten balloons.
‘What was most rewarding about this was how we demonstrated the enormous potential of human networking,’ says Toshiba Professor Alex (Sandy) Pentland, who heads the Human Dynamics group. ‘It was great that we won the contest, but more significantly, this exercise showed how building the proper incentives into a viral collaboration can quickly harness a large population to work together to address broad societal needs. It has helped us better understand how information spreads and why people cooperate.’