Hunch: The Search for Better Search
When Caterina Fake was picking out a new bag for her laptop computer, she didn't turn to friends, browse store aisles, or page through magazines for advice. Instead, she consulted her new Web site,
. In response, the tool asked her a series of questions, including "How much are you planning to spend?" and "Would you prefer leather?", before dishing out three recommendations.
Opened to the public on June 15 after a three-month preview for select users, Hunch is a tool for finding answers to a wide variety of questions—from mundane shopping decisions to dilemmas as serious as "Should I get a divorce?" The questions and answers are created by users, and the site uses feedback from the community to refine the relevance of results.
Hunch is the latest of several decision-making sites that have cropped up in recent years that are designed to shake up a Web-search landscape that has long been dominated by keyword queries. The sites, including Answers.com ( (ANSW)), , , , and Yahoo Answers ( (YHOO)), aim to do what Google and other keyword-based search engines have trouble doing: delivering results tailored to specific human situations and problems without sending users to many different sites on the Internet. "Google can't provide all the answers," says Danny Sullivan, editor-in-chief of Web site Search Engine Land.
Hunch co-founder Fake, who also co-founded image-sharing site Flickr and sold it to Yahoo! in 2005, isn't out to topple Google ( (GOOG)). But she does contend that in an online world where people frequently contribute edits to Wikipedia and thumbs-up stories on , the power of online crowds has yet to be fully harnessed to help people make better everyday decisions. "One little action is actually fairly insignificant in the grand scheme of things, but in the aggregate, these things together create something that's incredibly valuable," she says.
Fake: Flickr, Yahoo Answers, and now HunchFake burnished her search skills while still at Yahoo after the Flickr deal. She helped to create one of the most successful question-and-answer sites, Yahoo Answers. Launched in 2005, each month the site attracts tens of millions of visitors who pose questions and get rewarded with recognition for providing the most popular answers to others' questions. In fact, the site gets many of its visitors from traditional search engines like Google, since Yahoo Answers turns up in results when people search for questions like: "Does chocolate spoil?" Fake left Yahoo last year.
Hunch arrives at a response after asking users about five to 10 questions. Going a step beyond Yahoo's community-voting model, Hunch determines the best response for each individual user based on their past clicks. The site has voluntary survey questions on its front page, in the section "teach Hunch about you," where it asks simple, fun questions and then factors in those preferences when it's advising you on a decision.
Answer engines have a poor track record of late. In May, Microsoft ( (MSFT)) closed its MSN QnA, a service similar to Yahoo Answers. In March, Wikipedia founder Jimmy Wales announced he was closing Wikia Search, a short-lived project intended to create user-influenced results to keyword queries. Google once offered a collaborative search site called Google Answers, where people could offer money rewards for thoroughly researched responses to queries. The company axed the site in 2006, telling users "the Answers community's limited size and other product considerations made it more effective for us to focus our efforts on other ways to help our users find information."
A large base of active users is essential for making these types of sites useful, says Brady Forrest, a technical evangelist at who previously worked on the search team at Microsoft and helped develop the prototype of MSN QnA. "You have to constantly be able to pull in a certain crop of answerers who will invest the time to go in and talk about whether you should buy this laptop or that laptop and keep it updated," he says.
"they might get more valuable advertisers"Profits also hinge on these sites reaching a wide audience. "Many of these services are ad-based, which is problematic, says Greg Sterling, founding principal at technology market researcher Sterling Market Intelligence. '"In order to make that work, you have to have a massive audience.""
Could answer services one day be as lucrative as the search business? Eschewing banner or text ads, Hunch plans to get all of its revenue from fees it receives by sending users to shopping sites like Amazon.com ( (AMZN)). So while it may profit from only a small portion of the questions on its site, search experts say these links could be very profitable since people are already coming to the site with the intention to buy something. "Potentially, they might get more valuable advertisers," says Search Engine Land's Sullivan.
Sullivan says Google and other keyword search engines may lose a small portion of users to decision sites like Hunch, but generally he thinks the category is creating new searches that weren't done before at all. "They enlarge the search pie overall," he says, adding that when video started appearing on the Web, it prompted a similar boom in new types of searches that didn't exist before.
Thas isn't stopping traditional search engines from trying to compete with these upstarts by trying to offer smarter searches. In May, Microsoft unveiled Bing, a new incarnation of its Web search that's billed as the "decision engine." The site includes information such as average plane ticket prices for a search of "Chicago to London" right on the first page of search results.