The trackback URL for this entry is:

Trackbacks:

print this article

Crowdsourcing vs. distributed reporting - is there a difference?

Crowdsourcing has been in the news a lot lately, especially with Gannett’s plans to base its next generation newsroom structure in part on the concept. I’ve also read about distributed reporting, open source reporting and collaborative reporting.

Is there any difference between these terms?

For my money, distributed reporting relates to a specific and fixed-time project, answering a specific question. And by its nature, distributed reporting inherently includes some aspect of journalism since it is “reporting.” Asking readers about voting irregularities, for example, would be a form of distributed reporting since the need was reporting for a timely news story. For an example from the Cincinnati Enquirer, go here.

Crowdsourcing, meanwhile, is more like outsourcing, the term from which it was born. It aims to harness the power of community on a continuing basis to improve a service or information base. When we built an online map plotting all the places in our coverage area to go for free wireless Internet access, The News Tribune then asked the public to submit locations that we missed or that have since opened. We also invited them to comment on the locations and add photos, enhancing the original service. Dozens of readers have contributed in the past six months.

The concept of crowdsourcing might seem to lend itself especially well to grass-roots organizations and projects. But some of the most notable examples of crowdsourcing have come from some very big companies, including Proctor & Gamble, Amazon.com and Google. Check it out:

—Proctor & Gamble launched a Web site called InnoCentive offering some serious cash rewards to some 90,000 freelance scientists who can solve problems that the company’s 9,000 scientists can’t. It now works with other companies as a sort of crowdsourcing broker, allowing them to use the site to solve problems of their own.

—Amazon.com describes its Mechanical Turk project as “Artificial Artificial Intelligence.” It pays people to complete tasks that people do better than computers, such as identifying subjects in photographs and translating text. This is the opposite of the InnoCentive project. The pay is low and the tasks can be done by anyone. People need to perform a high volume of tasks to make any real money, but the tasks are so simple that some 10,000 people have registered to “turk.”

—Google doesn’t pay people to participate in its Image Labeler program, but it make the exercise so fun that it can be addicting. The goal is to improve the quality of Google’s Image search. Over a 90-second period, you are shown random images and asked to provide as many labels as possible. You “play” with another random user and when the two of you agree on a label, the software gets smarter.

Let me know if you see these emerging concepts differently.

Posted by MarkBriggs on Wednesday, November 15, 2006
(0) Trackbacks • Permalink