Flickr founder has Hunch about online searches
The woman behind Flickr has created what’s designed as a more personal alternative to traditional search engines. Hunch.com aims to give more relevant answers to queries by using questions to build up a profile of individual user tastes.
I have to admit that when I first heard about the site, and taking into account its look, the use of the phrase, “Today I’m making a decision about…”, it’s founder’s surname, and the timing of the publicity, I assumed it was a spoof of Microsoft’s Bing. Far from it.
Hunch works on the idea that the combined knowledge, experience and judgment of the masses is a source of knowledge greater than its parts. The site aims to help people make decisions rather than answer a question or find a site on a particular topic.
Before asking for help with a decision, users can answer questions about themselves on a variety of seemingly obscure subjects – when I tested the site I was asked whether I had my tonsils, what I thought of Alfred Hitchcock movies, what pizza toppings I liked, what type of shoes I preferred, and whether I knew what the Latin term for truth was. Users are then asked to rank the suggestion the site offers for the decision they are researching.
The theory is that the site can then build up a database of how which solutions are most popular with particular types of people. For example, Hunch already claims to have found a consistent pattern of people who vote a particular way in the U.S. preferring a particular brand of bottled water. The idea is that the more refined a picture the site can build up of a particular user, the better its final suggestion will be. Of course, the challenge for the site is that many people will avoid it until the solutions are more accurate, but that the solutions won’t become more accurate until more people uses it.
Caterina Fake, the founder of Hunch, stresses that the site is not a search engine. Speaking to CNN, she notes that people already use their social groups to draw on combined knowledge (for example, asking friends to recommend a yoga class). She notes that the internet already has much of the data needed to extend this process (for example, finding people who can recommend a yoga class in a foreign city you are visiting), but that there needs to be new ways to map out this information.
Or, as the Los Angeles Times succinctly explains, “Hunch computes answers by comparing what it knows about you to what it knows about people like you.”
Related Posts:
