Google Experimental Search: Building On Co-op?

Google’s existing dominance of the web rests directly upon the power of search. Most people honestly don’t know that Google’s search algorithm is based on a formula that includes text indexes (yes, the words on the pages matters), but also involves link count: the number of times people have ‘voted’ for the page by linking to it also matters.

In essence, the Google search machinery has always depended on the millions of individual links that people create in order to make sense of the Web. Now, they are involved in a new experimental project than might go much further in harnessing human grey matter directly into the architecture of search.

Google Experimental Search, originally uploaded by Stowe Boyd.

This experimental search project, a search page from which is displayed above, employs Digg-like voting on results, user reordering of results, and user recommendations for better pages. At this time, these actions have only local, user effects. However, one can extrapolate that in the future, individual results could be consolidated and analyzed, leading to a user-augmented search result.

Duncan Riley at Techcrunch says If you saw this one coming, give yourself a very large prize. Ok, I confess that I saw it coming.

As recently as a few weeks ago, working with a social search start-up, I suggested that their plans were dubious since it was inevitable that Google would incorporate more and more social information into search, and that no start-up could catch them. I more or less said the same thing to the folks at Hakia, when they were showing me their semantic search tool at the Web 2.0 Summit.

And I don’t think that this is out of left field, Duncan, since the fabric of the Web includes both the text analysis from indexing, and the links, which are purely social gestures left behind for Google’s spider to find. Also, Google Co-op has been around for a few years now, it is all about user customization of search.