Friday, November 03, 2006

Have You Heard of Crowdsourcing?

A post on Lorcan Dempsey's site a few weeks ago caught my attention. He drew attention to a concept defined by Wired Magazine writers Jeff Howe and Mark Robinson called crowdsourcing (Article). They defined the concept (to paraphrase from wikipedia) as activity traditionally completed by selectively hired, trained and managed workforces migrating to low paid or un-paid amateurs. These people use their knowledge and spare time to complete tasks, share ideas and solve problems. The obvious question is why don't you get what you pay for? The answer seems to be that to work well there needs to be a strong and widely held common purpose. The 'crowdsourcing' moniker is evidence that the idea is growing mainstream (and is itself a result of increasing wide-spread access to networks). As some have pointed out the Linux and Firefox development projects have been early examples of this concept.

The wikipedia site also lists some additional examples from mainstream companies such as Proctor and Gamble and Amazon. Recently, The Economist wrote about an Australian mining company that made available their prospecting documents under the guise of a competition. Interested 3rd parties could identify areas where they believed the company could mine for gold. Apparently the participating companies identified over 110 target sites of which 50 were new to the company. The benefit for the mining company was a significant reduction in time and expense to find these new targets and they also got access to new prospecting technology and processes.

Jeff Howe has a blog related to this topic named unsurprisingly The article in Wired is fascinating as it points to how business models have to change in some of the least likely businesses. Who would have thought that the stock photo business could collapse because we all have digital cameras and loads of images we can now share or license. Suddenly 'good enough' content exists in substantial amounts and the market is becoming over supplied - at least from the perspective of the stock photo agencies. The article points to as examples of the above and in the same vein exists to enable anyone who witnesses a newsworthy event to upload their photos for distribution and licensing.

Coincident with my reading Lorcan's blog entry about crowdsourcing I also witnessed an incredible example of this idea on the site. A few weeks ago, they decided that they wanted to translate their pages so that they could appeal to foreign language speakers. Instead of finding a language speaker in each of the target languages to slog through all the pages and translate them into F, I, G, S etc. they opened up the site to allow iterative translation. To my mind the results were astounding because they had for example 75% of the German translation done inside a day. To date, about a month after they started the project, they have 100% of the site translated into German, French, Dutch, Spanish and Portuguese. Here is the hit parade. As they (Tim) point out, adding a league table of translators seemed to incite some competition. There have been a number of blog entries on this topic but here is an early one.

Crowdsourcing is powerful, and yes we can argue it isn't that new, but the enablers are now more prevalent which is better and closer computing power, easier access to networking and rapid adoption of virtual social networks. My characterization of 'good enough' above was purposeful; one of the tenets of crowdsourcing is that the power of the network will outstrip that of a small group of experienced professionals. This is the real danger for publishers and others who have built silos of expensive (to develop and sell) content. As a database publisher I could point out the blemishes in 'good enough' products that I competed against. What happens when I can't do that. Witness the EB versus wikipedia debate earlier this year.

No comments: