Crowdsourcing: Why the Power of the Crowd Is Driving the Future of Business by Jeff Howe
With the success of Google, Social bookmarking, Wikipedia, and Open Source software it is readily apparent that leveraging large numbers of contributors efficiently can lead to substantial value.
Google captures the information contained in inbound links to assess the quality of the web pages that it has indexed, and thereby determines the relative ranking in search results. Social bookmarking provides the opportunity for browsers to vote on the individual pages that they visit, and draw potentially massive attention to what is important. Wikipedia has created an information source which exceeds the largest available 'traditional' encyclopedias by orders of magnitude, and open source software developers have created operating systems and applications suites which are superior to commercial software products. So, it is clear when you can leverage many contributors, you can achieve great advances.
Crowdsourcing cites many such examples. It is quite a lengthy book, and contains roughly the same message told several different ways, depending on the application area in which crowdsourcing finds deployment. In addition to Google and Wikipedia, the examples range from technological problems distributed on an eBay-like site for scientists and engineers, stock photograph marketing sites, designer tee-shirt sites, popularity contests, better software development through competition, and so on.
The examples are interesting. You will learn a lot about the ways in which entrepreneurs are currently successfully (iStockPhoto) and not so successfully (CurrentTV) going about exploiting crowdsourcing. However, a key aspect of Crowdsourcing appears to have been missed by the author. That is the efficiency of exploiting the input provided by the crowd. Google uses the resource of existing web pages to pick up on the selections made by millions of web masters in the links that they chose to create, the information is there to be gleaned as Google crawls the web. Exploiting that information is not trivial, it requires significant amounts of storage and indexing capabilities, and Google has developed the necessary infrastructure to make that feat possible. Likewise, the social book marking sites provide value to their users in storing the browsing favorites and provide that storage efficiently, in order to extract the value of the information that the millions of stored votes on individual pages provide. Wikipedia has developed a site capable of handling enormous volumes of contributors and reviewers, without the need for a significant permanent staff. The foundation of open source development has been attention to efficient build systems which allow diverse contributors to build software in the same way and provide inherent and efficient delocalization. In contrast, centralized nightly builds, upon which all developers depend are favored by older, centralized software development outfits.
I enjoyed learning of a range of examples of crowdsourcing from the book. I had not heard of the NetFlix effort to create optimal recommendations based on a competition on the web before, for example. I also enjoyed hearing about the fact that 'outliers' who tried crazy things, not obvious to experts, often drove crowdsourced problem solving in profitable directions.
The book ends with a collection of rules derived from the rest of the text these are something along the following lines:
1. Choose a model from: collective intelligence, creation, voting, or funding.
2. Choose participants to act as influence providers to usher the crowd.
3. Offer incentives and recognition to all participants.
4. Don't think of crowdsourcing as outsourcing.
5. Your crowds will need leaders who influence.
6. Divide your work into small simple tasks.
7. Be prepared for much chaff, as much as 90% chaff in fact.
8. Also be prepared for 10% or so of valuable output.
9. Your community will always be right.
10. Serve your crowd, do not oppose it.
As I mentioned, the clearly missing aspects of the book from my perspective were suggestions on the mechanics of achieving crowdsourcing. How do you go about decomposing a problem and providing an infrastructure that can be used by many people and not cost a significant amount in development or support?
For me this is a 4/5 read. A useful collection of examples and illustrations but lacking in practical guidance.



