Skip to content Skip to navigation

Research & Ideas

Search this site

Stanford computer scientists create crowdsourcing website to draw ... crowds

Unlike other websites that raise money, Catalyst is an experimental event-planning website designed to make organizing and participating easier than ever.

A game of chess in which Stanford students, dressed in black or white, are the pieces. The event was coordinated using Catalyst. | Justin Cheng

Stanford computer scientists have created a website to help organizers plan events that are more likely to succeed or allow them to pull the plug on impending flops before they occur.

The website, called Catalyst, is based on a behavioral science concept known as the threshold model of collective action, which posits that people may be reluctant to commit to participating in activity until they see others taking part, at which point interest surges and the activity becomes successful. But if participation doesn’t reach this threshold point, the event is likely to fail.

Catalyst builds this principle into software. The website allows people to enter a few details, such as date, time, description of the event and the number of participants needed to make it a success. If signups don’t hit this threshold point by the deadline, Catalyst emails organizers and would-be participants a warning.

Justin Cheng, the doctoral student in computer science who designed the site, says this fail-safe feature allows organizers to focus their energy.

“You can defer your investment until after you get a guaranteed amount of support,” he said.

According to Cheng’s faculty advisor, Michael Bernstein, an assistant professor of computer science, the goal of the project is to bring people together. “If we make failure safer, successes will be more numerous,” he said.

Cheng and Bernstein released a conference paper on Catalyst in October, including data from a three-month study of 30 trial events. Cheng will present this research at the Conference on Computer Supported Cooperative Work and Social Computing in February 2014.

In addition to making it easier to create events, Cheng and Bernstein wanted to engage more participants. So they made it easy to share events on social media. More importantly, perhaps, they gave organizers the ability to set roles for their events.

By allowing potential attendees to sign up for jobs with varying levels of commitment, they hope the site will entice “slacktivists” – people who vocally support a cause but don’t do anything – to attend events and participate.

The two think that, by combining their threshold model with easy-to-do roles, events will appeal to more users. “If people don’t have to do as much, they might become more likely to participate,” Cheng said.

Cheng and Bernstein developed Catalyst while looking at other crowdsourcing websites. The vast majority of these sites operate without users ever lifting more than a finger – people contribute money or digital signatures with mouse clicks. The Stanford team wanted to get Catalyst’s users working together in the real world.

“It’s not about giving people money, it’s about participating,” Cheng said. “You can only do so much online.”

So far Catalyst has been used by a food bank in Fremont to coordinate volunteers. Teaching assistants on campus have used it to schedule office hours. Stanford students even attempted to use Catalyst to organize flash mobs – semi-secret choreographed gatherings. About 46 percent of the events that garnered support on Catalyst ultimately succeeded. This is comparable to the 42 percent success rate of projects on, a crowdsourcing website that uses the threshold model to raise money rather than increase participation.

During Catalyst’s launch phase, users had difficulty using the site for certain types of events. Tri-City Volunteers, one of the largest food banks in Alameda County, found that it was not ready to handle the demands of coordinating more than 1,000 volunteers. Melissa Ponchard, executive director of Tri City Volunteers, added that new volunteers required training, adding complexity to their roles.

“[But] for nonprofits and volunteer groups with discrete needs, it’d be a really good thing,” Ponchard said.

Cheng agreed that smaller events, such as office hours and study sessions, were more successful. Such events might have a threshold point of only five participants and had an obvious direct benefit for participants, making more students willing to commit.

Usefulness to the participant was a key. Cheng noted that, although Catalyst had no enforcement mechanism, nearly 100 percent of participants who signed up for such events followed through.

But for purely social affairs, participation was lower.

In a human chess game in which students played as the pieces, 30 percent of the sign-ups did not attend. (Curious onlookers did join the participants who showed up, making the game a success.) Cheng suggested implementing a simple fix in the future: a one dollar deposit that is returned when a participant shows up to an event. Although it is a minuscule amount of money, he said, it makes their commitment feel more concrete.

“They feel a lot more engaged,” Cheng said, “and a lot more willing to contribute and participate.”

Matt Davenport is an editorial intern at the Stanford School of Engineering.

Media Contact

Tom Abate, Associate Director of Communications, Stanford Engineering,, 650-736-2245.