Do you need accurate sentiment for social media listening …

ByMike Moran chief, converseon, fellow, king, marketing, research, strategist, wikipedia Comments Off

Do you need accurate sentiment for social media listening?

■Crisis Management■Market Research■Reputation Management■Social Media Marketingby Mike Moran – Aug 27, 2012

Photo credit: Wikipedia

Most of you know that I serve as Chief Strategist at Converseon, so I spend a lot of time thinking about sentiment analysis techniques for social media listening. My background is in text analytics, so I am no stranger to technology solutions for this problem, but I’ve long felt that technology is only part of the solution to the vexing problem of social media sentiment analysis. The question is, how accurate do you need sentiment analysis to be? Sure we’d all want to know with 100% accuracy whether customer comments about our company are positive or negative. But are you willing to pay for it?

A few years ago, it was a silly question. If you weren’t willing to pay for a social media monitoring platform, then you wouldn’t get one, but HootSuite and other low-cost and no-cost competitors have jumped into the market. Those platforms don’t do what the high-end ones do, but they work just fine for simple use cases, such as crisis monitoring.

All these low-cost vendors have a problem, however. This race to the pricing bottom leads them to remove costs to lower price, which means that they can’t afford such fancy add-ons as sentiment analysis. Or at least no sentiment analysis that actually works.

And the market is accepting that for a very good reason. Not everyone needs accurate sentiment analysis. If only some of the records are correctly marked with negative sentiment, that is OK for crisis management, because a true crisis spawns dozens or hundreds of negative comments. You can get a lot of them wrong and still find more than enough to flag this situation as a crisis.

And what about all those other uses? People doing market research or customer service can’t afford to get a lot of them wrong. They need accurate sentiment. But if all the monitoring platforms are cutting features to cut costs, how can those customers be satisfied?

There’s a gaping hole in the market that needs to be filled. That’s what Converseon did with its Convey API. Instead of having to buy an entire monitoring platform, you can just buy accurate sentiment analysis. At least some people think it was a great idea. DataWeek named Converseon the top innovator in data mining of 2012, so maybe you should check out the Convey API, too.

About Mike Moran

Mike Moran has a unique blend of marketing and technology skills that he applies to raise return on investment for large marketing programs. Mike is a former IBM Distinguished Engineer and the Senior Strategist at Converseon, a leading social consultancy. Mike is the author of two books on digital marketing, an instructor at several leading universities, as well as a Senior Fellow at the Society of New Communications Research.