In June of 2009, Flickr cofounder Caterina Fake launched Hunch, a site designed to help people make decisions and find stuff that interested them. It did so in part by asking lots of questions on every imaginable topic–”Do you like the smell of Play-Doh?”–and I thought it was pretty neat at the time. But I haven’t been back often, and Hunch doesn’t seem to have become massively popular just yet.
Marshall Kirkpatrick of ReadWriteWeb is reporting on the site’s revamped version. It’s shifted the emphasis from decisions to recommendations, and is now focused on using the profile it builds about you from your answers to those silly little questions to suggest books, movies, TV shows, travel destinations, and other things you might like.
Here are some of Hunch’s “personalized recommendations” for me:
- I’ve been to Italy, New York, London, and Tokyo already and enjoyed them all, but they’re five of the most generic suggestions possible. And I live a few yards from San Francisco–I guess you could make the case either that it’s a lousy vacation destination for me or one of the best of all. But in either case, it’s safe to assume that I already know about it.
- It’s possible that some of the credit cards Hunch recommends for me are good options, but probably not the Citi MmtvU Platinum Select Visa Card for College Students. As far as I can tell, Hunch doesn’t know whether I’m a student or not, so I’m not sure why it’s on the list.
- I suspect I’m not alone in being well-acquainted with pizza, lasagna, fruit salad, hamburgers, and baked potatoes and already knowing what I think about them.
In other words, Hunch doesn’t seem to know enough about me to make recommendations that are truly personal. I’m sure that the problem is in part that I haven’t answered enough questions for Hunch to have deep insight into what makes me tick. So I’ve been busy answering more questions. Such as:
I’m sure that as I answer more questions, Hunch will give me smarter advice. But I can’t help but think that the service could give me better advice about travel destinations, credit cards, and dinner ideas if it simply asked me questions about those topics. (Hunch’s initial version let users create decision trees that asked straightforward questions, and while that feature is still there, it’s been deemphasized.)
In general, Hunch has a high cutey-cute factor. Without using the service much, I’ve earned 2,326 banjos. Which is wonderful news. But I have no idea what a banjo is or how they help me, or why some banjos look like light bulbs and apples…
I still like the idea of Hunch and many things about its user interface. I think I’ll answer a bunch more questions and see if I get recommendations that make sense and tell me something I didn’t already know. And if you try out the service, I’d love to know what you think.