But what I find most interesting about InnoCentive is its success rate. A study of InnoCentive led by researchers at Harvard Business School found that nearly 33 percent of the problems posted on the website were solved within the specified time frame. In other words, a disparate network of strangers and amateurs managed to solve problems that companies like Eli Lilly, General Electric and Procter and Gamble - companies with thousands of scientists and huge research budgets - had been unable to solve internally. Sometimes, the problems were solved within days.
By studying which particular problems were solved, and by tracking the efforts of "solvers" as they worked together in online "problem rooms," the Harvard researchers could see what, exactly, made some problems more solvable than others. The key was intellectual diversity. If a molecular biology problem just attracted molecular biologists, then it tended to remain intractable. But if that same problem managed to attract a molecular biologist, a biophysicist, an organic chemist and a statistician then, chances are, the problem tended to get solved. Even more interesting, perhaps, is that problem solvers were most effective at the margins of their fields of expertise. Chemists didn't solve chemistry problems: they solved molecular biology problems.
This makes me happy for a variety of reasons, but I'm posting it here to show that I really do believe that crowdsourcing can work, even after badmouthing Intrade the other day.