Point/Counterpoint: A Matter of Sentiment

As everybody with Internet connectivity well knows, social media sites such as Facebook, LinkedIn, and Twitter have become go-to sources for all the latest dope, personal and professional. In your business colleague's Twitter feed, you're just as likely to find a minute-by-minute recap of a keynote address at a major industry conference as you are news that the towels in his hotel room are paper thin and about as absorbent as a piece of granite.

Relative to the latter, in many respects social media sites have become emotional dumping grounds. In postings long or short, depending where on the Web you are, people pour out their agony and angst, likes and dislikes, pleasures and displeasures.

From a business perspective, you might think about this as more of the same, only a different avenue into your customer's psyche. Company or product Facebook sites aimed at reaching out and touching customers join a long list of similar, although less technologically advanced, initiatives companies have engaged in over the years. Mailed satisfaction surveys, countertop drop boxes, customer complaint hotlines, feedback@customer.com, Web forms -- these are but a few examples.

But the rise of social media as a business tool coincides with an increased interest in analytics and the fostering of data-driven decision-making. As a result, many companies are emphasizing the need to dig into what's being said by their customers out on the wild, wild Web.

And so we've seen growing interest in disciplines such as text analytics and sentiment analysis, with the goal being the ability to search out and analyze customer complaints or compliments generated in the free-form Web universe (as well as in similarly unstructured emails, PDFs, Word documents, PDFs, and so on). Text analytics, as discussed here, has become standard practice for many marketing departments as they study customer verbatim-responses from various sources, for example.

Such is the case at Oberweis Dairy, in Aurora, Ill. "Analysis of text records to understand customer complaints and quantify product issues is now just part of what we do," Bruce Bedford, vice president of marketing analytics and customer insight, tells me.

Sentiment analysis is similar but aimed particularly at monitoring emotions expressed on social sites, as AllAnalytics.com community editor Shawn Hessinger described in his blog, Getting Sentimental About Social Business Intelligence. A relatively new type of analytics, sentiment analysis is one of those topics that generates strong reactions among industry watchers.

We've heard of a handful of studies and actual use cases that show the effectiveness of sentiment analysis. Yet the naysayers are aplenty. As blogger Scott Koegler wrote in a recent AllAnalytics.com chat, "Combine anonymous users, no context, no consequences to comments. I can't buy that Twitter is anything but noise."

Another AllAnalytics.com community member had shared his equally strong feelings in a message board post. "Analytics is a powerful tool, and like the Force of Star Wars fame, it can be used for good and for bad. I guess it's up to who you talk to to define what's bad and what's good. But I think micro-targeting people based on their emotional responses (or instabilities) to sell them junk food, consumer electronics, and the like leans toward the bad," he wrote. "Then again," he added, "I don't get paid to market and sell junk food, consumer electronics, and the like."

With sentiment analysis being such a new and touchy topic, we asked two experts to weigh in with their thoughts in our latest Point/Counterpoint debate, available here. Joe Stanganelli, social media consultant, likes the approach and thinks companies will be able to derive good business intelligence from customer sentiments expressed in social networking content. Pierre DeBois, a Web analytics consultant with Zimana, is less enamored with sentiment analysis. Companies must realize they have a long haul ahead before they'll be able to enhance business strategy via sentiment analytics, he says.

How do you feel about sentiment analysis? Read the debate and share your thoughts on the message boards.

Beth Schultz, Editor in Chief

Beth Schultz has more than two decades of experience as an IT writer and editor.  Most recently, she brought her expertise to bear writing thought-provoking editorial and marketing materials on a variety of technology topics for leading IT publications and industry players.  Previously, she oversaw multimedia content development, writing and editing for special feature packages at Network World. In particular, she focused on advanced IT technology and its impact on business users and in so doing became a thought leader on the revolutionary changes remaking the corporate datacenter and enterprise IT architecture. Beth has a keen ability to identify business and technology trends, developing expertise through in-depth analysis and early adopter case studies. Over the years, she has earned more than a dozen national and regional editorial excellence awards for special issues from American Business Media, American Society of Business Press Editors, Folio.net, and others.

Midmarket Companies: Bring on the Big Data

The "big" in big data is no reflection of the size of the organization embracing its potential.

Push Yourself to New Analytical Discoveries

Take inspiration from Christopher Columbus as you pursue your analytical journeys.

Re: I'm not that sentimental
  • 8/12/2011 9:30:56 AM

While no doubt that is exactly what many companies will do, I don't buy the budgetary argument.  Most of these companies will balk at something valuable and game-changing like this, but pay a social media "ninja" $80-$120k/year to count Likes, Followers, and (if the ninja is especially qualified) clickthroughs and congratulate them on their numbers.

And this is the fundamental problem with the web analytics sector right now.  Most people aren't measuring what really matters, or even understand why the stuff that matters matters.

And the few that do kind of get it misunderstand the power or true nature of it.  Instead, they are, as you put it, merely looking to see how "hip and cool" people think they are.

My take: If you're going to have a budget for experimenting with analytics, do it right.  Do something that matters.

It's a great opportunity for companies in other regards too.  How great would that be for your brand if yours was among the first companies to master the implementation and application of linguistic sentiment analysis?

Re: I'm not that sentimental
  • 8/12/2011 9:16:27 AM

Joe -- I bet a lot of folks agree wholeheartedly with you that, yes, there is lots of value to be gained from playing around with sentiment analysis. But at what cost? If you've got limited people and budget resources, how high on the list should experimentation with sentiment analysis be? Maybe pretty high -- lots of companies do count on their customers' perception of them as hip and cool and being hip and cool these days most often means really "getting" social media and knowing how to  use it effectively. But maybe not so high -- I think a big chunk of companies will be lookers-on, letting others figure out -- ie, devote resources to -- this sentiment analysis and only jumping in after the experimentation settles down and proven values become more commonplace.

Let Academia, Facebook and Twitter worry about it.
  • 8/11/2011 11:13:40 PM

@Joe    Kudos to you for trying to defend a topic that many deem just not worth the effort. While I see your point(s), Sentimental analysis is best left in my opinion to academia or perhaps to companies directly involved in producing this type of data - Twitter, Facebook ...etc. I would not personally recommend main stream companies to invest too much time and effort into something that is just too subjective in my opinion to be of much worth.

Re: I'm not that sentimental
  • 8/11/2011 10:00:13 PM

Certainly agree with you there, Joe. (Have a quick peek at my response to your comment on your post.) The question is whether this is still an unproven experimental procedure or a ready-for-prime-time technique. I'd argue the former not the latter. When it elects a president or sells an ungodly amount of spaghetti sauce, I have no doubt we'll all be using it. Until then, I think we'll have to wait and see.

Re: I'm not that sentimental
  • 8/11/2011 4:23:52 PM

No question that there is still plenty of tooling around going on with linguistic sentiment analysis to improve it even more.  Nonetheless, there is value to be gained from trying it, using it, experimenting it, learning from it.

The more we implement and study this, the further it can take us.

Re: I'm not that sentimental
  • 8/11/2011 1:29:19 PM

Shawn, I absolutely agree that there's intriguing potential here. But for the time being at least I get hung up on the emotional element. At this point, if I were budgeting for analytics, sentiment analysis wouldn't be a top spend priority at this juncture.

I'm not that sentimental
  • 8/11/2011 1:22:44 PM


I said it over on Joe's post and I'll say it here. I'm interested in the idea of pulling semantic data from social media and think it has incredible potential. If analytics can be used to measure the emotional state of the communicator on Twitter/Facebook below their stated intentions, there may be incredible insight to be gained. This said, it seems a very big if. There are some major studies documenting the possibilities already, but, like my friend Pierre, who has the counterpoint here, I question the ability to fully implement it. Time will tell.