Wikipedia Academy - "When Peer Production Succeeds", Keynote by Benjamin Mako Hill

As Wikipedians have collaborated over the last decade to build the most comprehensive reference work in history, they have created, and documented, a new class of cooperative work. The academic world -- inspired by Wikipedia's success and its novel form -- have responded with thousands of academic papers about Wikipedia and its processes.
But as many people who have tried to create their own wikis can attest to, Wikipedia's enormous success at mobilizing volunteers and building works of high quality has been difficult to replicate. Even the Wikimedia Foundation and Wikipedia's founders have struggled to replicate Wikipedia's success in their other projects.
Through the knowledge and intuitions of Wikipedians, and through the work of the academics studying Wikipedia, we know an enormous amount about *how* Wikipedia works. However, the fact that even our own community has systematically failed to replicate our greatest success points to the fact that we still know very little about *why* Wikipedia works. The situation is similar in regards to free software and other peer production communities. And although each failure can teach us something, we have not given these failures the same attention we've saved for the successes.
In this talk, I will refer to research on free software and free culture communities and suggest that the ideal of peer production is only rarely realized. I will show how free software, and free culture, only very rarely looks like its poster children: the Linux kernels and the Wikipedias. I will present some of my research comparing failed free culture projects to successes to both suggest a methodology, and a potential set of answers, in order to answering the question: Why did peer production projects like Wikipedia work?
I will suggest, and try to show, that by learning from our failures, instead of ignoring or sweeping them under the rug, we can make both free culture advocacy and free culture practice more effective.