Google looking glassA week ago I wrote a piece here called Google Search Winners and Losers (and what does it mean for you?) The primary point I was making was that seeking to play SEO games (search engine optimization) with an author site was a waste of time, money and energy. While I supported the point with some Alexa charts for several book-oriented Web sites, that’s hardly empirical proof.  Today, I happened to stumble on something that further supports that point, but and may indicate something even more intriguing.Here’s what happened: on April 12 I posted a review here of a New Yorker piece by Ian Frazier, and in that review, I compared Frazier’s writing favorably to that of John McPhee. This morning, while replying to a comment someone posted to that blog entry, I needed to remember the name of one of McPhee’s books, and ran a Google search of  “John McPhee.”  The result? The third hit of the search – out of 9,260,000 hits – was my blog entry.  Huh?

So I tried “Ian Frazier,” and got the same result – it was the third hit, this time out of 3,530,000.  Huh? again. It was then that I noticed that both links weren’t to the blog entry itself, but to the Google+ outtake take I’d posted.

There are at least three possible explanations for these results. One is that Google’s algorithm is doing some subtle things like giving higher rankings to recent posts (very likely, and not a new technique, and perhaps in fact this posts rank will fade) and perhaps even using semantic search techniques to recognize a review as a review and assigning hits of this type a higher presumptive value to searchers. And indeed it’s no secret that Google has made substantial investments in incorporating such techniques into its search technology.

The second is the possible relevance of the tens of thousands of pages I’ve written at this site: http://www.consortiuminfo.org/ and the fact that there are tens of thousands of third-party links back to those pages, which presumably gives my name some credibility in the Google algorithm.  And in fact there are quite a few searches you can run (e.g., “consortium standards”) that will generate first search page hits to pages at that site.

The third, of course, is that Google might be biasing search results to favor its own social network. Now wouldn’t that be interesting?

Google, I’m sure, would strenuously deny that that’s the case. On the other hand, the same searches didn’t produce any hits on the first several pages at either Yahoo or Bing. To be fair, the overall results were startlingly different, so clearly the algorithms used at each site vary considerably in other ways as well.

But wait, it gets a bit weirder.

Someone else tried the same search, and didn’t get the same result. Huh? What could explain that? Incontrovertibly, Google must be somehow tailoring search results to the individual, applying the same approach to search that it does to targeting advertising to the pages you view. But based on what? Recent searches (as with advertising)? Other information it is keeping about the individual (shades of NSA)? So I tried the same search on the same computer, but using the Chrome browser – and sure enough I also got a different result.

What’s troubling about this experience is that however Google generated this result, it rendered a less useful result. It even raises the possibility that this type of customization could lead a user to more narrow, or even biased, results. Do we really need a Cable News approach to search?

I’ll leave to you which conclusions you wish to draw from this little experiment.  But given the fact that the search took me to the Google+ squib rather than the blog entry itself, it couldn’t hurt to post the occasional link to your writing at Google+.

Have you discovered The Alexandria Project?