Google moving toward personalized / behavioral search

Reading Nomeeja today, i found a post about the new Google Frequent Searcher Program. The “program” is currently just a number: using a cookie, they count the times a person searches on Google and displays a counter showing the numeric count and a colored bar at the bottom of the Google home page.

Neither NoMeeJa nor Clickz cites the reason behind the new feature, they only mention the official Googlespeak, that there is no winning in it for the user, it’s solely for self-awareness. Well, i don’t buy that.

In an article i read while researching the future trends of search engine marketing for my thesis, i came across a mention of the possibilities of “personalized search”. The idea is that Google (or other search engines) would move beyond the algorythm method and introduce a behavioral and artificial intelligence aspect into the delivery of its search results, trying to guess what the searcher’s intentions and goals are with their search, based on patterns emerging from previous searches done by that user.

The frequent search program could be the first step in that direction. With the cookie Google is now tracking its searchers individually. Of course it’s very basic and has limitations (e.g. the cookie can be deleted/disabled, if several people use a computer, it counts them all as one person, and if you use several computers, you will have a different counter on all of them), but it is a first, baby step.

Another aspect and a different route towards personalized search is demographical personalization. There’s a Clickz article by Danny Sullivan, that explores the subject. Different people may be entering the same query in the same search engine but anticipate different results. Gender, age, nationality, geographical location all have a significant impact in what they want to see. The examples from the article: flowers and football. When a man searches for “Flowers” they want the websites of flower delivery companies. If a woman enters the same word into a search engine, they will more likely want gardening tips, seed merchants, and the like. (These are results from a 1999 study by a personalization company.) Similarly, a european searchiing for football will find Google’s current #1 result, entirely irrelevant, while #2, will likely be what he’s searching for (and vice versa for a US searcher). It would be interesting to find more such examples (please post in the comments here if you have ideas!). (Local search is another emerging trend in search engines, and is also a first step into the direction of personalized results.)

In addition to the new frequent search program (it’s not generally available yet, only testing!), Google recently acquired a company called Outride, who developed technology that uses surfing behavior as a means to deliver personalized results.

My first prediction for 2004 is that Google will continue moving into this direction, taking the step from global relevancy to individually personalized relevancy.

Comments are closed.