If the term "automated sentiment" sounds to you like something out of science fiction, you might be rather new to the world of brand management. Rest assured the term has nothing to do with progress in the field of robotics. Automated sentiment is a way of gleaning information about public opinion without anyone actually having to read any of the public's opinions.
What Is It?
Automated sentiment assessments are being offered more and more by firms that specialize in social media monitoring and other brand strategy tools as a way to map what consumers are thinking about your company. Firms develop and use algorithms that sort through the language of reviews, tweets and other online commentary to give you a summary of opinions, whether positive, negative or neutral.
Some platforms get much more detailed than that, breaking down very specific information about the source of the positive, negative or neutral comments, whether blogs, tweets or other sources. They can help you find trends in the way people are talking about your business and create a timeline so you can connect certain events to tides of opinion. They can also help you gather this information about your competitors so you will know better how to position yourself for victory.
But Does It Work?
The accuracy of automated sentiment is a topic of much debate. Most of the conversation on the topic online reflects a lot of skepticism. No one really believes that a computer can be perfectly accurate when summarizing human emotions. Most humans cannot even do that. But research shows that automated sentiment analysis is usually about 70 percent accurate, which could be a lot worse. It is difficult for an algorithm to detect sarcasm, for instance. One blogger told of his social monitoring tool's evaluation of this comment: "To be honest I like my Droid way better than my old iPhone." The platform deemed that comment as positive for iPhones, even though it is clearly not. Possibly the automated sentiment tool picked up on the word "like," which is usually positive. Clearly there is some weakness in the system.
Evaluating the Value
Because the accuracy of automated sentiment is constantly in question, the value is too. Many in the blogosphere have voiced opinions on either side. Some feel it is completely worthless. We disagree. Though automated sentiment ratings are not completely accurate, the tool makes sense if you have a large volume of public opinion to analyze. Getting it right about 70 percent of the time is better than 0 percent for sure, and as long as you don't hang your entire brand strategy on automated sentiment ratings, you will do just fine.
Automated sentiment is one tool that can be useful in first analyzing and then improving perception of your product. As often as you can you should engage humans in the process of finding out what people think of your brand. There is really no substitute for having someone read posts, reviews and other comments that refer to your company. When that is not possible the use of an automated sentiment application might provide some helpful insight.
At TopTenREVIEWS We Do the Research So You Don’t Have To.™