Prof. Rosen doesn't actually give a bulleted list, but a collection of best practices seems to emerge from his examples. The underlying message, is that crowdsourcing is just a method of data collection. Expect results to be data, not an answer to your original problem.
Here's my crack at sorting the speech into a coherent checklist:
Crowdsourcing Checklist Inspired by Jay Rosen
- Open your data. If the data needs to be collected, expect to open source it as it's gathered.
- A community must have a common problem. The community doesn't have to be geographic, but it helps.
- Divide the problem into small, well defined, tasks that can be filled out on a form and done in minutes.
- Publicize. Use an established platform to ask for help, and keep asking.
- Have a timeline. Tweak as the presentation as necessary to meet the goal.
- Continuously coalesce the data. The result is data. It needs to be compiled into a human-readable format. This should be an ongoing and public process to let the community know how close they are to reaching the goal.
Those are the lessons I gleaned from Prof Rosen's talk. I plan on seeing how they fair against, projects like today's SETI announcement that they would release their data to the public. It seems that their brute-force approach of creating one of the world's largest super computers didn't work out. They're hoping that open sourcing the problem of literally finding signal in noise will find the extraterrestrial life for them.
SETI was originally following the checklist pretty well (they didn't have a clear timeline, but it's a little hard to blame them for that). It didn't work out for them, we'll see if this new approach – which seems to ignore the list
– is any more successful.