Flickr founder goes with Hunch for decision engine start-up

The premise is a simple one, users ask Hunch a question and it gives an answer based on their unique personality type, sharing advice on everything from the best car to buy to whether its wise to get a tattoo or not.

Hunch knows who's asking the question based on its own questions, asked in a personality-type survey. When first logging on to Hunch, users are presented with a first question, and after answering anywhere from 10 to 1,500 of these, their Hunch profile is built.

Questions Hunch asks its users range from broad idealogical queries, like 'Do you tend to support liberal or conservative policitians?' to the bizarre 'Do you enjoy the smell of Play-Doh?'.

In turn, users can then ask Hunch from some answers, for example, when asking 'Should I buy a Blackberry or an iPhone?', in which Hunch asks a few more questions, such as 'Do you want to watch videos on this device?' before given a definite answer.

Hunch bills itself as the anti-search engine, but curiously uses the same 'decision' strapline as Microsoft's Bing. But instead of scouring the web for links and keywords, Hunch finds what users want by what it knows about them, and those similar to them.

The more the website learns about each individual personality, the better it customises results, said Fake, adding "it's like a friend getting to know someone's taste and preferences over time, so they can provide sound and trusted advice".

The website has been in development for about a year, building up a base of personality types and question-answer fields duing that time. It currently offers advice on over 2,400 topics, with over 14,000 follow-up questions.

Fake, who was also instrumental in launching Yahoo Answers, said it had some initial affiliate partnerships to help monetise the website, for example directing users to third-parties, such as Amazon, after finding out what kind camera was best for them.

Fake said Hunch will not sell any personal information to brands or advertisers.

Hunch is also social media inclined, with a large community base of users who can share advice or ask their own questions. The website is also looking at developers to create an API that would combine Hunch and Twitter in some capacity.

Based on user feedback, Hunch is giving the right answer about 80% of the time, said Fake, but the company wants that figure to be higher, in the 90-95% range.

Hunch has 10 employees, based in New York and recently raised $2m in venture capital funding from Bessemer Venture Partners and General Catalyst Partners.

This article was first published on revolutionmagazine.com

Only registered users may comment. Sign in now or register for free.