Google Universal at SES San Jose 2007
As this was the first major search event following the launch of Google Universal, no surprises about which topic would really dominate this year’s conference. Let’s say I didn’t see any spare seats in any of the sessions on Universal, 3D, Blended or any of the other terms devised to describe the new search interfaces. Not having a unifying name did seem to be a problem, if only to provide a simple point of reference in conversation at the bar. I guess once this new generation of search interface has aged sufficiently, we’ll all be happy to continue calling it just ‘search’.
My interest however was fixated firmly on Google Universal. Don’t get me wrong, I like Ask 3D and happy to acknowledge that most engines have been doing some level of blended results for some time. But, I still think Google’s Universal was the most radical and technically challenging.
Determining relevance for vertical results over regular results is a huge undertaking. The factors that determine reputation for text based content are very different to those used for videos. So comparing apples to pears is a challenge that can only be done subjectively. This would certainly explain the strange results observed in the UK as universal rolled out. For a while company sites were demoted over news articles even for a brand search. Even highly commercial searches were infected with news articles where user intent was clearly abandoned.
I was lucky enough to have had the opportunity to talk to David Taylor product lead at Google, just as Universal was being rolled out in the UK. When questioned about these anomalies, he was keen to defend the robustness of the product. This is not some beta test, but a system designed to get better over time. I can only imagine that once a result is deemed worthy for inclusion in regular results, they can monitor user behavior to adjust the ranking in real time. This seems to fit in with our observation – those strange results only stayed for a couple of days. Though there are many still popping up.
An important question for me and most search marketers was the blending of Universal with Personalisation. David was unwilling to commit here, but thought it was reasonable to assume this could be a natural progression. Even the possibility of combining group behavior is something that can not be ruled out looking at his Amazon roots.
One aspect of Universal that I find the most remarkable is the ability to determine the relevance of a news item in real time. Matt Cutts confirmed that Google can determine a new news item worthy of inclusion within regular results in less than an hour. It can’t be too long before we’re going to see this real time context evaluation applied not just to news results but regular web page content.
Did you enjoy this post? Why not leave a comment below and continue the conversation, or subscribe to my feed and get articles like this delivered automatically to your feed reader.
Comments
No comments yet.
Leave a comment