Google May Get Personal

With last month’s acquisition of Outride, Google may be poising itself to go into an area of search refinement that no major player has gone successfully before: personalized search results.

With personalized results, a person would get back a list of results that takes into account some of his or her demographics. For example, if you had registered yourself as a man, you might see a different set of results when searching for flowers than a woman might see.

Indeed, the above differences aren’t just speculation but reality. When Direct Hit experimented with personalized results back in 1999, it found that in a search for “flowers,” men typically wanted sites that let them send flowers (no doubt because they’ve done something wrong, I always like to joke, when telling this story).

In contrast, women often wanted sites that let them order flower seeds or plants for gardening purposes (which is pretty typical of my wife, who does a lot of gardening and who also never does anything wrong necessitating that she send me flowers).

Similarly, knowing your age might let a search engine deliver more relevant results for music-related searches. Understanding where you live can be helpful, too. For example, those in the UK searching for “football” almost always have no interest in information about American football and instead want sites that deal with what North Americans would call “soccer.”

I’ve long thought that personalization would be the next big technological solution toward producing better Web-wide search results. Basic text retrieval was greatly enhanced by the use of click-through measurements and link analysis. Personalization seemed about to happen in 1999. Direct Hit was poised to move forward with it, while another company in the space, GlobalBrain, was acquired by NBCi (which was then called Snap).

Sadly, nothing came of these developments. NBCi died, having never implemented the personalization elements of GlobalBrain’s technology. As for Direct Hit, further testing found that users were afraid that personalized results would mean they might miss something, because the technology might incorrectly presume to know what they like or dislike. Concerns over privacy are another reason why personalization hasn’t made inroads with the major players.

Personalization did seem like it might get a second chance through development by two new companies that gained some attention in mid-2000: Buzz Notes and GroupFire. They aimed to follow your surfing behavior as a means to deliver personalized results.

GroupFire, a spinoff from the Xerox Palo Alto Research Center (PARC), even announced partnerships with Excite@Home and Inktomi in December 2000. That was when the company renamed itself Outride. That was also the last thing we heard from Outride, until now.

As of last month, Google has purchased Outride. There are no plans to introduce any components of Outride’s technology overnight, but the search engine does expect to make use of what it has acquired, in some way.

“There’s nothing immediately ready to go, and there are no concrete plans I can share with you right now,” said Sergey Brin, Google’s president of technology. “They are one of a dozen types of technologies we might use.”

Google didn’t disclose a value on the Outride acquisition, but Brin did say that it was “a smaller meal” than its acquisition of the Deja.com archives last February, the value of which also remains undisclosed.

Overall, don’t expect personalized results to appear in the immediate future at Google. Indeed, they may not come at all. However, the acquisition probably gives the concept a better shot at life than it has had for some time.