Human capital: First, we had an amazing 'long view' leader who ran the show. He saw the project at a higher level of abstraction, could understand the details but didn't get bogged down in them. The work was both motivating and inspiring, this intrinsic motivation is critical to finishing a project of this scope (4 years in duration).

Requires dedication and sacrifice from the key participants, all of which is much easier to gain when you're working on a consciousness raising activity that gets people up in the morning.

Required deep negotiation to collaborate with data providers, having someone who can drive these talks is key. A mental map of the data landscape is similarly important: where can you use public data, and reuse what has already been done?

Requires you can learn new skills as you go along and perform tasks well outside your area of expertise. Constant learning isn't optional.

Filling a need: On many levels, but key is understanding how we effect our planet. Having a clear way to communicate your idea mattered, in our case a map: http://www.nytimes.com/interactive/2008/02/25/science/earth/20080225_COAST_GRAPHIC.html

Getting the job done:

Your initial idea is nothing like the final result: the seed question turned out to be the wrong one, and the project shifted in response. Ideas are cheap, but good implementations are expensive.

Scaling is critical, constantly amassing new data to improve the analysis, and tuning what you're measuring. Running the model once after optimization took 18 hours, we scaled horizontally to compute Monte Carlo simulations.

Much guidance provided by highly experienced participants early on in the project, helping shape it and form critical networking links, similar to the role played by YC. Rely on existing relationships when you need help, and look to others who can help. Going it alone has merit, but rely on existing relationships for help when you can. Ultimately what matters is getting the job done.

Positive feedback is a strong motivator, which is good as high levels of motivation are necessary to beat out both internal and external obstacles. Capital is limited: this is the economic law of gravity, bootstrap or climb your way out of it.

Where it differs:

Do everything yourself: would be nigh impossible without heavy collaboration and synthetic effort, blazing your own trail wouldn't get you to the same endpoint. Strong bilateral agreements are the only thing that flies here.

Release early, release often: while internal updates to the group were critical, the system doesn't reward openness early in a project, and the public face is nothing till the publication date.

The business model: in a startup, you have hard numbers to measure returns. Science measures returns in indirect ways, and the time scale for realizing those returns is generally much longer. Traffic isn't a good metric for this project, but is interesting: in two weeks, our website has seen 30k uniques, with another 22k uniques to the KML file. These results were mostly due to the press coverage from big players such as NYT, NPR, NatGeo, MSNBC and a few social media sites.

Our project website: http://www.nceas.ucsb.edu/GlobalMarine