Caterina Fake’s Hunch Is Now a Personal Recommendation Service, What’s the Question?

Hunch has relaunched as a more clearly defined “taste-driven” Net recommendation service. It’s a clever move, since as the Net penetrates deeper into our lives, we may find ourselves reaching to systems like this for help.



Hunch, the brainchild of Flickr co-founder Catarina Fake, has adjusted its web service so that it’s now more clearly defined as a recommendation service, rather than a social Q&A service which the folks at ReadWriteWeb note many media outlets have mistaken it for.

Hunch describes itself as a system to “offer you a great recommendation to address your choice, problem or dilemma, on thousands of topics” based on “the collective knowledge of the entire Hunch community, narrowed down to people like you, or just enough like you that you might be mistaken for each other in a dark room.”

How does it work? At its core is a matching algorithm that measures how much your Hunch profile matches up to similar people who use the service, borne of thinking by its founders that machine learning “could be used to guide practical, smart, and highly customized recommendations.” To use it you hook it up to your Twitter or Facebook account, and then answer an initial batch of just 20 simple multiple-choice preference questions–you can choose to answer more, but 20 will do. For example, building my Hunch profile I discovered that just 39% of Hunch users are male, 36% consider themselves extroverts, 67% have never been whisked to hospital in an ambulance, 25% have ridden in a helicopter and a shockingly-low 36% have changed a spark plug. Hunch uses these answers, in a dynamic and adaptive, learning way to then decide which books you may like, which European cities you should visit…and so on. The site provides these automatically, but there’s a search box for you to prompt it to recommend something to you. Check my recommendations out in the image up there: It definitely worked out I’d like some Python and Fawlty towers, and further down the list it was pretty good at guessing which classical composers I’d like (Bach as number one).

What Hunch has done in its redesign is put these recommendations at the heart of its offering, and though other parts of the system are still accessible (like the eerily accurate Twitter prediction game, where Hunch pre-guesses how you’ll answer questions) they’re via sub-menus. Why did it choose to do this? The answer is something similar to why iPad apps like news aggregator Pulse and Flipboard, or Twitter’s trending topics exist: As the Net grows, and extends its tentacles deeper into our everyday lives, so its influence on our decisions will grow, as will the sheer number of pieces of information available. Tools like Hunch are only going to get more and more common, as they intelligently digest down information on the Web into useful chunks that you can much more easily manage.

To keep up with this news, follow me, Kit Eaton, on Twitter.

About the author

I'm covering the science/tech/generally-exciting-and-innovative beat for Fast Company. Follow me on Twitter, or Google+ and you'll hear tons of interesting stuff, I promise.