Many of you may have heard about or read the book by James Surowiecki, The Wisdom of Crowds. If not, the brief explanation is that one attempts to use the collective wisdom of many to optimize a process or decision (Wikipedia definition). This idea is nice in theory, but what is the result in practice? I'm sure you've heard the old phrase: "Too many cooks spoil the broth". Is this concern valid? Does crowd wisdom optimize a result?
I decided to test an actual example using search engine results. NASA has implemented a "crowd wisdom" search engine. In brief, NASA's web pages are tagged (by the system, not the user), and as the user interacts with the web site, the visitors actions are recorded and a complex algorithm decides what web content you found most engaging ... and compares that to your original search terms (read the NASA search engine case study ... via Baynote)
Here is my test ... Google Search vs. NASA Crowd Wisdom Search for the Orion Nebula.
- Google Search Query: Orion Nebula (query restricted to NASA.gov)
- NASA.gov Search Query: Orion Nebula
You can click upon either search query and examine the results. You'll notice that the NASA.gov search results display how many users utilized a certain result, and that there are suggested follow-up links in the left hand margin. All of these data points result from the fact that NASA.gov's content is tagged, and user interaction is recorded and evaluated. Search results rankings on NASA.gov are constantly changing as users vote with their feet (or in this case ... mouse clicks). Once again, see the NASA case study to learn more about this tagging process. My Google results are just the "normal" results using Google's index.
I have included two screenshots of the top five results from each search (click to maximize). You'll note that their is very little overlap in what content is served up to the user for the top 5 results.
Now you could dismiss the fact that I'm only showing five results, and comparing ten results. However, consider this study from Cornell University (via Online Marketing) ... and your own personal interaction with search results. Once you get below the first THREE results, very few users actually click through to content (remember SEO?! ... that's another story and blog post). (click upon the image shown below to view full sized).
I hope I've made you think, and perhaps consider how the wisdom of crowds (i.e. social interaction with web content) may be used to optimize content ... even without the user taking any conscious action (i.e. page content that is tagged by the system, and then having the user's interaction with that content evaluated).