Hunched Over

Last week I got a note that Hunch is scheduled for a June 15 public launch. I hadn't been very active on the site these past couple of weeks -- too busy binging it. When I logged back in today to see what I'd been missing, I had this notification waiting for me that the topic I submitted, "Will Hunch be bigger than Google one day?" was declined. Ruh roh. Can't be a good omen...

OK, now that I've gotten the obligatory snarkiness out of the way (truth is, it really wasn't a real decision I submitted anyway so the community was right to reject it), I can proceed with my review. Having spent the last 30 minutes playing around with Hunch (hunching it? getting hunched? satisfying my hunches?) I must say I'm impressed how far it's come since I first tried it. The community has certainly seeded some interesting threads allowing the "decision-engine" to learn and evolve. Clearly, Hunch gets better the more its used.

According to Hunch, more than 40,000 people have used it during this preview stage. And, on average, they've answered more than 120 questions about themselves. There's no doubt teaching Hunch about yourself is the best way to derive value from it. Here are some of the topics that were recommended for me...


Given how heavily the Bing vs. Google debate has been weighing on my mind, I decided to "play" the topic, "Which search engine should I use?" Sure enough, this topic has been played 178 times since being created on March 30th. Here's the qualifying question I was asked...

I selected general engine and, bing, my decision was made...


Why did Hunch pick this? Good question. Fortunately, there's a link for that...

There's also a list of pro's and con's for the decision as outlined by other Hunch-seekers like me. (They even have little ratings next to the usernames to show how closely their preferences match mine, presumably so I can decide how reliable their comments are.)

And therein lies the beauty of this platform. Notice there were no keywords or queries here. Just a nice linear progression and results presented based on a combination of indexed information from the web, my personal preferences and crowd-sourced data. Of course, Hunch will not always get it right. (But "right" is subjective anyway, right?) There is a built-in feedback mechanism so it can learn and most of the other "decisions" I've tested Hunch on ask more than one qualifying question before returning an answer. (I guess Google was just SO obvious!)

Here's a more practical example -- "Where should I eat in Chicago?" I'll spare you all the screens and just share the "Why did Hunch pick this?" from the end as it shows the qualifying Q's that led up to this answer...

Seems to me like you could take Bing's first TV commercial and swap Hunch in. Hunch is not about keywords and queries. (The only time keywords are used is to map the question you've asked to others that have already been "played.") And it's not about spouting off a bunch of related information sending you down a wormhole of links and facts. It's about helping you make a decision.

As much of an upgrade for Microsoft as Bing is (to the point where it's barely recognizable from Google) it's not really a decision engine. Heck, it's not even a search engine. I like how David Berkowitz characterized it as a search portal.

Is Hunch the real decision engine? Wait 'til June 15th and decide for yourself.