• by David Berkowitz , April 11, 2006

Long live personalized search; personalized search is dead. The same can be said of social search and vertical search. All three are melding, and a search giant has an opportunity to bring this hybrid technology to the masses.

Let's start with the personalization corner of the triangle. I use all the major portals' personalization services, and I'm toying around with some new entrants into the space, such as Findory for personalized news, Windows Live, and Claria's PersonalWeb. These sites almost always direct me to content I would not have found otherwise. Personalization is thus an always-on complement to word of mouth marketing. The technology becomes a more efficient and perhaps even smarter version of that friend or nephew of yours who keeps sending you all the best links.

All of this makes me wonder about the prospects for social search. Here, that nephew and a number of your peers are connected through some network, and all of your search results are affected by what others in the network search for. Thus, networks of fishing buddies, jazz musicians, beer aficionados, and shoe lovers will all find different search results coming up for the query "bass."

Yahoo has integrated social search components into its My Web 2.0, a tagging and bookmarking tool. My Web users conducting searches will often find links atop the natural results to relevant pages that other My Web users bookmarked. While this may allow some lower-ranking pages to rise to the top, the results aren't necessarily more relevant this way. The people who are doing all this tagging don't necessarily have anything to do with you and your interests.

Eurekster has refined its version of social search into the concept of "swickis," in which a blogger or site owner can offer a topic-specific search engine that matches its results to the site content and other searches. The value proposition, however, is strongest for the publisher, as consumers, in the short term, are still conditioned to turn to actual search engines for their searches.

To share one final example, Friendster, powered by Eurekster and Yahoo Search, allows users to see what people in their networks are searching for. Yet this is confined to searches within Friendster, and Friendster isn't a search engine. At most, you use Friendster to search for your friends. As for my friends, they're desperate. Three of the top four searches in my network are "true love quiz," "handsome men," and "how to win an ex back." The other one? "Halloween costume." Either my friends really hate paying retail, or Friendster gave up on updating this six months ago. So who's in the best position to do social search right? Social search will go mainstream when it's incorporated into a top-tier search engine's main results. Consider how it would play out with Google, as an example.

For Google, social search would involve tying a few pieces together of what it's already doing. Google already personalizes results for registered users based on their previous searches. It could further personalize search results based on groups of registered users-- namely, those on a user's Google Talk list. Now, following this line of logic, assume that this group of registered users is clustered largely around one or two common themes, such as technologists in the Pacific Northwest or Bible Belt NASCAR fans or Binghamton University college students. With personalized search results based on a presumably like-minded community of users, it all starts to resemble vertical search, a concept Eurekster acknowledges about its swickis.

To recap, personalized search becomes social search. Social search becomes vertical search. All of this with the goal of making search results more relevant.

Yet how much relevance will users be able to stomach? How much will users crave hyper-targeted media at the expense of the communal experience? As personal as the technology gets, the watercooler effect isn't going away. However, people may redefine the watercooler. Instead of caring what 30 million people are reading or watching or listening to, people will start caring about, say, the 150 people in their first degree and the few thousand others one degree further out.

If this environment takes hold, with the watercooler evolving into the nexus of who you're in contact with and are likely to come in contact with, then the stage is set perfectly for the emergence of personal-social-vertical search. The biggest question that remains is whether the technology will arise to support the cultural shift, or whether the shift will occur thanks entirely to the technology making it possible.