Seven Breakthrough Sentiment Analysis Scenarios

Brand Reputation Management is scenario six, another application of listening (monitoring, text analytics, presentation) technologies to online and social media. These are places where anyone -- past/current/prospective customer, self-styled know-it-all, industry authority -- can post opinions that can damage or boost your reputation.

Brand Reputation Management (BRM) is more a subspecies of marketing and than of customer experience management. It is product and company rather than customer focused. It reacts to and attempts to shape perceptions rather than to manage experiences. Sentiment analysis is an essential component.

Online Commerce is the final breakout application of sentiment analysis that I will describe. The assumption is that consumers value others' opinions, about travel, restaurants, stores, and products, opinions presented both in narrative form and reduced to aggregate ratings. Opinions guide us in searches, hence Bing- and Google-computed star ratings (as described, for instance, in the Googler paper Building a Sentiment Summarizer for Local Service Reviews). Opinions also help us find products and content that appeal to us once we do reach a site. Note the centrality of reviews at sites such as Amazon.com. Here, look to vendors such as OpenText that provide both sentiment analysis and the semantic navigation capabilities (such as faceted search) needed to exploit sentiment classifications. (Disclosure: OpenText is a consulting client of mine and a Sentiment Analysis Symposium sponsor.) Opinions influence purchases. Look for sentiment analysis built-in to become a standard feature of online commerce sites. Let the Analysis Begin I've described seven sentiment-analysis scenarios, seven areas where automated analysis of attitudes and opinions is shaping voter, customer service, marketing, consumer, and company images and actions. The business value of attitudes, opinion, and emotions means sentiment can't be ignored. The volume and velocity of subjectivity in enterprise, online, and social content means non-automated analyses are no longer, on their own, competitive. And I'll admit that I find applying machines to the task of deciphering human subjectivity simply fascinating. I think we will soon discover that sentiment applications are like the travelers of an old nursery rhyme: As I was going to St. Ives,
I met a man with seven wives
Each wife had seven sacks
Each sack had seven cats
Each cat had seven kits
Kits, cats, sacks, wives
How many were going to St Ives? There's a huge scope and variety of subjective information out there to be analyzed. That analysis will help us make better decisions about corporate, personal, and public matters. I expect we will find that my seven scenarios explode into a multitude of real-world sentiment use cases: No puzzle there. Seth Grimes is an analytics strategist with Washington DC-based Alta Plana Corporation. He chairs the Sentiment Analysis Symposium, slated for April 12 in New York.