The BLG Blog

Posts & articles that have helped thousands build performance through pragmatic leadership.

I Like Sentiment Analysis?

No matter how inane or frivolous your last tweet was, someone out there thinks it’s pretty important.  It used to be that when a company wanted to get the pulse of “the man/woman on the street” they went to elaborate, expensive and time intensive lengths to do so.  Well, thanks to the modern miracle of something called Sentiment Analysis, one man’s constant stream of trite, annoying and far too personal blathering is another man’s gold.

Sentiment Analysis, also known as the less confusing/more ominous sounding Data Mining, is defined as “the computational study of sentiments and opinions expressed in unstructured text documents”. In layman’s terms, companies take things like your tweets, Facebook posts, blogs and customer reviews and then change the way they do business to fit your mood. The best part of the whole thing? They don’t have to pay the social networker a dime. No focus groups, which means no moderator, no conference room and no free pizza and soda (my favorite part of the focus group experience.)  Of course this isn’t a free ride, in order to wade through the mountains of text, and all those pesky 01011010 it cost money, but in the end it’s cheaper, faster and once again, no pizzas.

Of course there are some downsides to this 21st century tool. Matt Rhodes from compared old fashion human analysis with the seven leading automated social media monitoring tools (Alterian, Brandwatch, Biz360, Neilsen Buzzmetrics, Radian6Scoutlabs and Sysomos). What he found was that when dealing with “facts of information” automated monitoring tools were as or more accurate than human analysis.  But when it came to actually deciding if a comment was positive, negative or neutral the automated tools fall far behind.  Rhodes points out that this is important to businesses because if you are reacting to and acting upon a large volume of negative comments that turn out to be positive (or vice versa) your entire operation can be compromised. Computers might be able to beat you at Jeopardy, but they still are having a problem deciding if you really loved that Nic Cage movie or you were in fact just being the snarky. And seeing that Twitter is about 95% snark, this is not a small problem.

Is this problem of perception a major stumbling block? If Mr. Rhodes is correct than yes, but with the way software and hardware are updated in the blink of an eye I am guessing this won’t be a problem for much longer. But for now you need to still be wary of any technology that claims to take away the need for at least some old fashion one on one human contact with consumers. Plus, I still want a shot at that sweet free pizza.



About BLG

Whether you want BLG to deliver a complete leadership academy across your organization, focus on a key group, or supplement an ongoing program, the primary goal of any of our programs will be business impact.

Recent Posts


Sign up for our Newsletter