Hunch Decision Engine Takes On Google, Bing

Hunch, the widely anticipated new crowdsourcing project from Yahoo Flickr co-founder Caterina Fake, launched to the general public in the form of a decision engine to help users, well, make decisions.

Hunch, which will inevitably be helped by Microsoft’s plan to throw $80 million in marketing dollars toward Bing, its erstwhile Google-killing decision engine, is geared to help users make choices based on information they provide. Hunch relies on crowdsourcing, building on the collective participation of its users (currently 40,000-plus from the beta version).

While Bing is still a search engine at heart like Google—letting users enter keywords and queries to find information on a variety of topics—Hunch culls information from users by asking them questions about topics. Hunch explains in its FAQ page:

“When Hunch proposes a decision result, it will also show you why it proposed what it did. If you disagree with some of the reasoning, you can correct it. If you think Hunch missed asking a crucial question, you can submit one. And if you think Hunch is missing a good result, you can add that, too. Hunch collects and organizes all this input so that it becomes smarter for the next user.”

For example, to find out whether or not one would be interested in working at a startup company versus an entrenched player with brand solidity, Hunch asks the user a series of related questions to help the user make a decision.

Specific questions for this startup topic include: Do you want to work somewhere that encourages newcomers to question the status quo?

Users can choose: “Yes, I want to be able to influence how the company goes about things,” or “No, I prefer to work where there is a well-established way of doing things,” or skip the question if they opt to.

Yes or no questions are common, but there are also multiple-choice questions. Hunch weighs the user’s answers in total and, based on how a user responds, concludes whether or not he or he would be happy at a startup or more comfortable at an established company.

There are some 2,500 topics and counting. The more users forge new topics or participate in existing ones, the smarter Hunch gets because it builds on the answers participants provide it. The more people participate, the better informed the engine’s offered choices will be, which will in turn makes decision-making better for a user.

So how does Hunch make money? Basically by serving as an e-commerce referral system. According to the Hunch FAQ, certain decision result pages on Hunch link to external sites, such as Amazon.com or Best Buy, where users can purchase the product or service that Hunch proposed.

Should users buy the suggested product or service, Hunch may earn a referral fee from the merchant. But fear not the obvious potential payola shenanigans: Hunch claims the presence of a link to a retailer has no effect on the decision outcomes Hunch proposes.

This is much more of a decision engine than Microsoft’s Bing; Hunch will no doubt benefit from Microsoft’s use of the phrase decision engine to describe its revamped Live Search. The real test will be whether or not Bing, Fake’s evangelism and old-fashioned word of mouth exposure will help Hunch blossom.

Fake said in a blog post that, starting today, users can use Hunch without logging in, “though of course Hunch’s hunches are much better if you do.”

Hunch will also give bloggers code to post widgets. There’s also a new Explore page, which Hunch will keep updated with new topics, new users and “those interesting correlations that are the byproduct of all our question-answering technology.”

“Five years ago, maybe even three years ago, we couldn’t build a product like Hunch,” Fake told Search Engine Land’s Matt McGee. “Hunch had its way paved by Wikipedia and Yahoo Answers. It’s become more acceptable” to have crowdsourced, collective knowledge Websites.

Perhaps, but will users lend a hand to help Hunch tap the wisdom of crowds? As McGee notes, the Hunch experience is more gradual than other Q&A engines, building on the crowds and rendering information iteratively, like a Wikipedia page.

The real question is whether something like Hunch will ever be ported to enterprise scenarios. Couldn’t businesses benefit from Hunch’s approach to making recommendations to users?