Revolutionary search engine #2394: Hunch
From a Flickr founder, new search engine Hunch is out of private beta today.
So if you’re Binged out, you’ve been Google Squared and got your kosher-search on, you might need a bit of a helping hand with your next search. Hunch promises to help you find what you’re looking for on the web in “ten questions or less.” Huh. But in a searchy world where Google Answers, Wikia Search and MSN QnA have all disappeared from the landscape, will users trust the collective opinion?
Caterina Fake thinks so. Fake co-founded Flickr and had something to do with Yahoo! Answers before moving on to create Hunch. As you use it, Hunch asks a variety of seemingly irrelevant questions about you- are alien abductions real? Who makes the best fries? Do you like bumper cars? Building on your answers, the site compiles a demographic profile to find the best search results when you need them.
Clearly the site needs more input to work as intended. Earlier today, as I answered the random questions (not geared to a specific search) it would indicate that X% of users responded with that answer. But if it was a choice of three answers, my result was always the same as 33% of respondents, with no variance.
So I took the direct approach, selecting a question from the dropdown menu at the upper right hand corner of the page. Typing in “is” lead to a bunch of questions- I chose “is my partner cheating on me?” (Sorry, honey, it’s for work!)
A series of questions followed- does he tell you where he’s going? Do you both say “I love you?” Does he let you in on his schedule? After the promised ten questions, I was awarded with this result:
I don’t think Hunch will replace Google, nor is it looking to. But it might provide a comfy home for the legions of search engine surfers seeking answers to more personal questions like should I marry my boyfriend or is my son gay?
Hunch currently has a user satisfaction rate of 80%, and Fake believes with more users, it can go as high as 90-95%. So, what do you think? Will you be using Hunch?
