Hunch Inc. Fact Sheet

Thanks for your interest in Hunch!

Updated February 1, 2010

What is Hunch?

Hunch is a recommendation site that gets smarter the more it's used.

Hunch will propose a concrete and customized recommendation for thousands of topics of every kind. Hunch refines its results by asking you questions about both the topic itself and who you are. What kind of car should I buy? Should I switch to a Mac? Should I dump my boyfriend? Where should I go on vacation? Should I get a tattoo?

Hunch uses machine learning to get smarter in two ways:

  • User contributions train Hunch to be smarter overall. Contributions can take many forms, from correcting a fact that Hunch got wrong, to suggesting new topics to feature, follow-up questions to ask or recommendation outcomes to propose.
  • The more Hunch learns about each individual user's personality and preferences, the better Hunch can customize recommendations for that user. It's like a friend getting to know someone's taste and preferences over time, so they can provide sound and trusted advice.

What problem does Hunch solve?

Our long-term goal is for a user to be able to come to Hunch looking for a smart recommendation on just about any topic, and after answering a handful of questions, get as good a recommendation as if she had interviewed a group of knowledgeable people or done hours of careful research online.

Eventually, when Hunch gets good enough, we hope users will trust its recommendations without having to turn to lots of external time-consuming sources of information.

What stage is Hunch in?

Hunch opened to the public on June 15, 2009.

Hunch currently offers more than 6,000 topics, 30,000 follow-up questions for those topics, and more than 50,000 possible recommendation outcomes.

  • As more people use Hunch, it will become both smarter about what it knows and broader in the scope of topics available.

How does Hunch make money?

  • Some of the recommendation pages on Hunch link to external sites where users can purchase the product or service that Hunch proposed. If they do, Hunch may earn a referral fee from the merchant.
  • The presence of a link to a retailer has no effect on the recommendation which Hunch proposes. Within a given topic, it's likely that some recommendation pages will link to an online retailer, and others won't. Some topics don't have these sorts of links at all.

How is Hunch different from other sites?

  • Hunch proposes customized recommendations of every kind.
  • Hunch gets smarter by harnessing the collective knowledge of the entire Hunch community.
  • For each topic, Hunch asks a series of structured follow-up questions that lead to a concrete recommendation.

What's the history of Hunch and who is behind it?

  • Hunch is based in New York City.
  • Hunch has a 12 person team. The core algorithm was built by MIT computer scientists with backgrounds in machine learning. Hunch's design and product functionality are led by Flickr Co-founder Caterina Fake, who is a Hunch Co-founder and Chief Product Officer.
  • Hunch began a preview to the public on March 27, 2009, and opened fully to the public on June 15, 2009.
  • Wikipedia founder Jimmy Wales joined Hunch's Board of Directors in late 2009.

Contact for additional questions

Kelly Ford
VP of Marketing