Businesses want to extract value—relevant and actionable data—from the expansive pool of content that the social sphere provides. By evaluating the opinions that have been expressed—from the very positive to the very negative to every other relevant feeling or statement in-between businesses can identify trends and specific examples that can be turned into insight for decision-making. The aspect of opinion mining that has received the most attention is sentiment analysis, which is primarily used to track user attitudes about a specific topic or the opinions of a certain population.
Using technology to derive a “sentiment score” from the opinions expressed through user-generated content online can be extremely useful for organizations in evaluating a large data set of social brand mentions. It also can provide a straightforward way to segment and filter content based on positive or negative commentary. Sentiment scoring allows businesses to isolate the themes or issues driving consumer sentiment, and enables dynamic and illustrative reporting of trends and market reactions—or time-sensitive and reputation-threatening situations such as product recalls.
However, one of the key challenges in understanding and applying a sentiment analysis solution in the business setting is that sentiment is not a one-dimensional result with a universally agreed-upon set of criteria—and this is particularly true when evaluating social media content. Sentiment is subjective, exists in varying degrees, relies heavily on context (meaning, it is not just about positive versus negative), and can be measured at different granularities—from the document, paragraph, sentence, phrase or pattern level, or some combination of those.
In our recent white paper Measuring Social Sentiment: Assessing and Scoring Opinion in Social Media, we discuss why it is critical for businesses to understand that while powerful and accurate scientific methods can be applied to analyzing what consumers are expressing online, truly understanding the context, intent, tone and humor that is inherent in human communication is both an art and a science. And when evaluating social sentiment analysis you should consider the following two questions:
- What are my social media goals and how does sentiment fit into the equation?
- What type of reporting do I want to do and how does sentiment help me do that?
For instance, assess how deep your organization needs to go with its sentiment classification. Most solutions tend to divide sentiment into three classes: Positive, Negative or Neutral (no sentiment expressed). Others offer a fourth option for sentiment classification and scoring – Mixed (both positive and negative in a single post). Would it be beneficial to your business to know if consumers had mixed feelings about your brand? That insight might make all the difference in helping to nudge an “on the fence” consumer successfully toward making a purchase and ultimately becoming a loyal customer.
It is important to remember that sentiment analysis also allows you to go well beyond if the product sold to give you a connection to the customer and an understanding of their emotions and how they actually feel about your brand. It also allows you to isolate the themes or issues driving that sentiment so you can better react to the market and drive higher levels of brand affinity.
Author: Jackie Kmetz is director of data strategy and product training at Visible Technologies, a leading provider of Social Intelligence solutions designed to improve and accelerate business outcomes.