Steve Mansfield operates his own Internet search engine from a place he calls a "secret hideout" — a small office surrounded by low-rent apartments on the outskirts of this college town known for its horse farms.
Mansfield conceived Prefound.com a few years ago on the premise that humans, from pretty much anywhere, can collectively provide better intelligence than a computer program developed out of the Silicon Valley.
Other startups, too, have had similar visions for "social search." And today, even large competitors like Yahoo Inc. and Google Inc. are pursuing the concept, hoping it'll help make search results more meaningful and thus expand the companies' market share.
Traditional search results are largely based on objective criteria such as counting the number of links other sites have placed to a given Web page. Social search gives people subjective answers — the best sushi restaurant in Chicago or the best Web site for information about French impressionism — not necessarily the site visited the most.
"You're essentially breaking up a problem and sending it out to a huge number of people for a query, getting answers back," said Steven Jones, a communications professor at the University of Illinois at Chicago. "It kind of ends up being greater than the sum of its parts. Other people are going to make associations and connections to information you probably would not have made."
At Prefound, launched earlier this year, users contribute to the knowledge pool by submitting clusters of sites they believe would appeal to like-minded people. As an incentive, the largest contributors even get a share of Prefound's advertising money.
A visitor looking for information on, say, New Jersey beaches can get the user-recommended sites, grouped by users. One user's cluster gives you restaurants, Internet cafes and other information on the coastal town of Ventnor City, N.J.
Results are better the more people contribute sites.
Big players are paying attention
Jones said it's too early to know whether social search will dramatically change the way people look for information on the Internet, but it's already changing the way traditional search companies do business.
Yahoo, a distant second to Google, has entered the game largely by buying some of these startups, namely Del.icio.us, a system for discovering new sites based on shared bookmarks, and Flickr, a photo-sharing sites where users tag items with keywords to help friends and strangers alike discover photographs on any topic.
Google has started to incorporate community answers on travel and health questions into its main search engine. It has also established a program allowing users to contribute their own content, tagged with specific attributes, to turn up in search results.
"To some extent the small companies have invented it, but the big companies have been thinking about it for quite a while, too," said Charlene Li, an analyst at Forrester Research.
Steven Marder, co-founder of Eurekster Inc., considered one of the earliest social search sites, said Yahoo's and Google's entry into social search was "validating our philosophy and methodology."
While the change in direction at the Internet search leaders proves the startups were onto something, it also is forcing them to either find a specialized niche soon or get swallowed by the much larger fish.
Marder said Eurekster, where results are weighted based on how many users click to a given site, will never be a destination site like Google or Yahoo, but he is trying to market the service for companies that want to build their own specialized search engines on private Web pages.
Likewise, Prefound is largely trying to cater to academics.
Other startup efforts include the appropriately named StumbleUpon, which three Canadians designed to cater to habitual Web surfers. Type in a topic and click "Stumble" to randomly be diverted to a site popular with other users.
"It's more of a recommendation engine than a search engine," said Garrett Camp, one of StumbleUpon's founders. "All they really want to do is discover all the best sites up there. Google is still going to remain focused on the task-oriented. StumbleUpon is much more discovery."
Such is the mantra for many of these startups that have seen Google or Yahoo upstage them on the concept of social search. They're marketing themselves as more vital than ever, hoping to pick off a few users, which might be all they need to turn a profit.
"There's room for lots and lots of players in the search race," said Chris Sherman, executive editor of Search Engine Watch. "We've got the major players out there, but as people get better at learning how to navigate the Internet, they're not going to necessarily be looking for an answer from these titans anymore. They're going to be looking for more specialized or personalized information."
Early search engines largely looked at keywords embedded in Web pages, but site owners could contaminate results by tagging their sites with as many words as they could conceive.
Google then came along with page ranking, essentially allowing the community — through Web site owners — to vote on the relevance of a site. The more other sites link to it, the more it is deemed worthy. But site owners, particularly those with big budgets, wind up with the greatest influence in the results.
Let the community decide
Social search attempts to let the entire community of users decide. In theory, the best ideas win out in the online marketplace and are less open to manipulation.
The approach is not limited to search. Cloudmark Inc. uses its base of users to judge what is spam and what isn't. News aggregation sites like Digg Inc.'s digg.com and Time Warner Inc.'s Netscape.com ask visitors to recommend and vote on items to go on top.
Among the larger search companies, Google recently launched Google Co-op, an information-sharing feature that marks its first major foray into social search.
Although the site isn't fully operational, Google users can see some of the results when they search the main engine for information about health issues or travel. The results appear as special links at the top.
Shashi Seth, product manager for Google Co-op, said there will always be a role for traditional search algorithms, and he doesn't see specialization ever taking away much business from the major companies.
"They're all very niche plays in the area of search," Seth said. "We still believe people will find core search fulfilling the bulk of their needs."
But Eckart Walther, vice president of product management for Yahoo, said that although traditional search technology can be refined, it will likely remain best for fact-based searches. Social search, he said, provides the promise of answers to more subjective questions.
Yahoo's MyWeb ranks Web sites, such as favorite hotels, based largely on how many people choose to bookmark them. Del.icio.us does a similar job, but users rather than software determine what keywords should correspond with which sites.
And Yahoo Answers allows people to write specific questions that other users or field experts then answer. Microsoft Corp. is working on a similar program, and Google has one in which it charges a fee.
"If you asked me what the best hotel in San Francisco is, we're different people, have different friends and different opinions," Walther said. "Search engines are terrible at answering these kinds of questions."
Mansfield acknowledges that Prefound and others like it will only get a tiny fraction of the pie in the search business, but that could be enough to turn a profit. After all, they know what they're up against.
"The board of Google could publish their laundry list and get into The New York Times," Mansfield said. "That's quite a hard thing to compete with."