Wednesday, May 30, 2007

Word of the day: crowdsourcing

Apparently this word is about a year old — and I'm amazed that I hadn't heard it before now. I came across it in Tapscott and Williams (2006) Wikinomics, which I mentioned in my list of new cooperation books. This May 27, 2006 article from Jeff Howe's blog documents the word's origin. It starts out by saying that 9 days earlier Google returned only 3 hits for crowdsourcing. At the time of the writing Google returned 182,000 hits! From 3 to 182,000 in a little more than a week! (This was a year ago, and it was only a year later that I first heard the word.) This follow-up article completes the history of the term.

Wikipedia's Crowdsourcing article (appropriately) captures the meaning.
Crowdsourcing is a neologism for a business model in which a company or institution takes a job traditionally performed by a designated agent (usually an employee) and outsources it to an undefined, generally large group of people in the form of an open call over the Internet. The work is compensated with little or no pay in most cases. However, in a few examples the labor is well-compensated. In almost every case crowdsourcing relies on amateurs or volunteers working in their spare time to create content, solve problems, or even do corporate R&D.
When Debora and I talked about it she raised the issue of prize money offered for accomplishment. Are the various DARPA Grand Challenges crowdsourcing? Was the reward offered for the first one-person non-stop flight over the Atlantic crowdsourcing?

Some of the uses of the term seem to include those cases as when a corporation offers money for the solution of a problem. The more general and intuitive use doesn't. Wikipedia is a crowdsourced encyclopedia produced by an entire crowd and not just by one individual or team from a crowd. Like Wikipedia, prediction markets are a crowd-based way of producing a result — in this case evaluating an issue. The entire crowd participates.

In some work that I'm doing, I claim that distillation mechanisms will become increasingly important. By this I mean mechanisms that filter out the essence or value from crowd generated raw material. Wikipedia has one distillation mechanism. In fact, its distillation review mechanisms is also crowd sourced. So Wikipedia is doubly crowdsourced: both in the generation and distillation of information. The academic peer review process is a less crowd sourced variant of this. The original input is crowdsourced. The distillation process is more formal — although also dependent on a crowd of reviewers.

Prediction markets are another distillation mechanism. The market mechanism itself is the distillation process, which is not crowdsourced. It is a way of extracting a single essence from the opinion of many people. Of course markets have been around for a very long time.

A third form of distillation mechanism is the offering of a reward as in the DARPA grand challenge or when a corporation offers a reward for solving a problem. Many people are encouraged to participate and produce a result that satisfies the stated need. Only one of these gets the prize. The decision about which participant gets the prize is made by the offerer, not by a crowd.

One of the oldest distillation mechanisms is the marketplace, both economic and more generic. Like Wikipedia most marketplaces use a two-level crowdsourcing mechanism. At the generation level, anyone has the ability to create and market a new product or service. At the distillation level, the crowd itself is the arbiter of what succeeds. The same is true for the Internet. Anyone can create a new website; the crowd decides which websites develop a following. What's different about the Internet is that the long tail effect allows sites with relatively minimal followings to prosper.

Another distillation mechanism is used in the ESP Game and PeekaBoom in which games are used to get people to contribute knowldge. These were created by Luis van Ahn.

Yet another distillation mechanism isn't really a distillation mechanism. Linus Torvalds talked about his GIT source control management system. It is a distributed system in that there is no central branch. Everyone has his own branch. The important branches are the ones that people decide are important. The SCM system doesn't tell you. It is agnostic with respect to which branch is the main branch. There is no main branch.

No comments: