Note:  Hunch (featured in this blog post) is the project of Chris Dixon, who also writes a fantastic blog, that I follow daily.  I’ve included a link on the left that allows you to discuss this blog.  If you haven’t already tried Hunch, you might start with “What Blog Should I Read?”

I had an interesting discussion over the weekend about whether social media is “just a fad”.  My argument was that the value of social media is not the novelty to the end user.  In my opinion, companies that are built around how enjoyable social media can be, are missing the point.  The true value (from a societal rather than individual perspective) is the data that can be derived from all of that tagging, tweeting, liking, digging, friending and sharing.  This data is what we call “Big Data” (Big Data is a much larger category, but the data abstracted from social media certainly applies).  Big Data is difficult to parse and hard to understand, but once you find a way to make sense of it the results can be both surprising and accurate.

Let me supply an example, Hunch.com.  Hunch uses formulas to make recommendations for particular “plays”.  Example plays include, “What Color Suit Should I Wear?”, “What should I do with my old laptop?” and “What should I eat for dinner?”.  Hunch asks you a series of questions for each play and then gives you a recommendation.  Let’s take the “What Color Suit Should I Wear?” example.  I told Hunch about the times I was planning to wear the suit (at work not at cocktail parties), the season I’d be wearing it (all year), the color shirts I like, and other such information.  The recommendation came up Gray (if your curious, I’m actually planning to wear a black and white check suit today).  I like gray suits though and thought the recommendation was appropriate.

“So what?” you say, any Kaufman’s sales rep could have asked the same questions and given me the same answer.  True, but in this case they didn’t have to.  You see some crafty user of Hunch defined the list of questions they thought would be necessary to determine the color suit I want.  Other people added additional questions or answer choices.  Then several people (including me) used the “play”.  Once we got our results we told Hunch whether we liked or didn’t like its decisions.  Lastly, and here’s the tricky part, Hunch then learned from the experience.  It learned what answer choices have the highest correlation to people who like particular colors.  The more people who run the “play”, the better it will get at creating a proper recommendation.  It may also identify surprising trends that your Kaufman’s salesman might miss (e.g. what if every person who wears red shirts, likes black suits?).  Hunch will notice that, and when you select red shirts, it will always recommend black suits.  What Hunch really needs to make such recommendations is a TON of data.  Fortunately, users of social media happily volunteer it.

There are a number of companies that already use this type of a concept.  Google search results are based on what other people clicked on.  Amazon recommendations are based on what people with buying histories like yours bought next.  When Facebook recommends friends or things to be a fan of, it recommends friends/things that your friends like.  There are also a number of exciting possibilities here.  What if WebMD kept track of everyone’s symptoms and what disease they ended up having.  That would be a pretty powerful predictive tool (though I’m not suggesting anyone throw out their doctor).  The bottom line is, social media’s primary societal benefit is not to give you fun gadgets, it’s to learn more about human nature by watching you.

The privacy debate is a whole other question, that I don’t have the space for here…