The word "sentiment," according to Dictionary.com, means "an attitude toward something; regard; opinion." OK, so here's my opinion on sentiment analysis: Its utility for marketing professionals woefully lags behind how people use social media.
In other words, apply a big dose of caution in how you apply value to the sentiments expressed on social sites. We have a long way to go before meaningful analysis can enhance business strategy with a deep understanding of customer sentiment toward a product or service.
It's not that sentimental value doesn't have its place, particularly from an analytics perspective. Sentimental value requires measurement, after all. But its measurement methodology differs from what we've used for Websites so far.
But consumer sentiment appears on the algorithms of social media platforms as well as on Web pages. With current social media analytics tools approaching message exposure in different ways, capturing the value underlying a comment includes some unrefined assumptions. You can deduce a visitor's impression of a Website from measurements such as average time on the site and number of page views. But does a word shown in a social media feed really constitute an "impression" that truly influences followers?
Another challenge is this: The algorithms in a social media feed rely on simple terms to display a person's sentiment, yet the linguistic definition of sentiment implies complex thought behind the expression. So how does a marketer get to the complex thoughts that lead to an action?
A great example is a recent Twitter trend on the phrase "White Diamonds," which became popular when Elizabeth Taylor died. But did that trend translate into sales for White Diamonds, her fragrance for women?
I did a quick Google search. While the term trended during the news reports of Taylor's death, I saw no indication of an associated increase in sales for the perfume. In fact, I discovered White Diamonds was already the top-selling celebrity-endorsed fragrance in the world. So what lift could be expected from sentiment when you are at the top of the game?
Now, you could correctly argue that data has no value without its human interpretation. But beauty is in the eye of the beholder, and the training and insight of the given beholder in this case, the business intelligence analyst, is still playing catch-up to the fluid changes of the Internet. Even the technology is catching up. Just recently, Google and Webtrends updated their solutions with enhanced social media measurement capability, for example.
Dictionary.com offers another definition for sentiment: a thought influenced by or proceeding from feeling or emotion. But can 140 characters truly reflect one's thoughts?
A number of events and products stir a quick emotion, such as a song or a sports team. But for an engineered product, such as a car, you just aren't going to capture why people like it in a quick Tweet. And the more passionate someone is about such a product, especially if it has a strong community built up around it, the more detailed the explanation will be. Just ask an Apple MacBook Pro or a Porsche 911 owner.
In addition, you must account for cultural and contextual interpretation in analyzing expressed sentiments. Social media are offered on the go, rather than simply in front of a computer, so a marketer would have to deduce the context influencing the sentiment of a Tweet or series of Facebook comments, for example. Is the consumer sitting on a stalled train and venting about all things, products included, because he's harried? How can you tell?
Furthermore, getting to an understanding of context requires knowing where the consumer is located at the time of the Tweet. While the consumer must permit use of location information, this could tread into a negative sea of privacy management that can turn off consumers and make a sentimental value campaign a bust. To quote rap group Run DMC, when a consumer uses the word "bad," perhaps it's "not 'bad' meaning 'bad' but 'bad' meaning 'good.' " It's also tricky to get context.
How ready are we to assess sentimental value? As they say in Twitterland, #notwellatall.