Text analytics vs. sentiment analysis

Chocolate and peanut butter. Romeo and Juliet. Yin and yang. Sometimes two ideas become so closely identified with each other that it can be hard to remember that they are actually separate entities. Text analytics and sentiment analysis make up one such pair. They are both ways to derive meaning from customer data, and they are both critical components of a successful customer experience management program. However, they are not the same thing.

Text analytics is the process of analyzing unstructured text, extracting relevant information, and transforming it into useful business intelligence.

Sentiment analysis determines if an expression is positive, negative, or neutral, and to what degree.

In other words, text analytics studies the face value of the words, including the grammar and the relationships among the words. Simply put, text analytics gives you the meaning. Sentiment analysis gives you insight into the emotion behind the words.

What's the difference?

They identify different kinds of content

Text analytics shows you what is being written about most. You can see which topics are trending, which ideas are commonly linked in the text, and even determine who is bringing up which subjects the most. When you apply sentiment analysis to that same content, it tells you if the topics are being addressed positively or negatively.  Sentiment analysis can also be applied to non-text feedback such as video, audio, and images – because someone smiling at you is giving you a higher sentiment score than someone shaking their fist in your direction.

They provide different kinds of early warning

Text analytics can provide a heads-up that trouble is coming when a new topic appears in your data. For example, if the word “spoiled” suddenly spikes in your restaurant chain’s feedback, you should look into that area quickly. Sentiment scores also identify potential risks. In this case, a decline in sentiment score indicates that some aspect of your business has left your customers feeling negative toward you.

They work differently

Text analytics from Clarabridge relies on our patented Natural Language Processing (NLP) technology to process text-based data in much the same way as the human brain processes language. Using proprietary algorithms it identifies the parts of speech, understands which words and ideas are linked, automatically corrects for mistakes, and derives meaning. It can comprehend the patterns and trends in a whole database, or drill down to understand a single tweet.

Sentiment analysis focuses on the meanings of the words and phrases and how positive or negative they are. Clarabridge gauges sentiment on an 11-point scale, which provides a more nuanced view of sentiment than the traditional “positive-neutral-negative” choices common in manual sentiment coding. For example, see this sentence, below:

I loved that laptop but this sale should have been easy and it wasn’t.

While someone applying a three-level scoring system would have to decide whether to weigh the love of the laptop more heavily than the difficulty of the sale to determine the overall sentiment, the Clarabridge sentiment analysis scale allows us to break the sentence down more specifically. The phrase “loved the laptop” can garner a +3 score, while “should have been easy and it wasn’t” gets -4. So while the sentiment of the sentence overall is negative, the two topics can be analyzed separately for a more accurate view of the customer’s feelings.

The Clarabridge CX Intelligence Platform applies both text analytics and sentiment analysis to feedback, in preparation for categorization and reporting. Combining the two types of analysis reveals the deepest, most specific insights that can be used to make bold business moves. Together, text analytics and sentiment analysis reveal both the what and the why in customer feedback. They are different, but they are better together.

For a detailed look at the technology powering Clarabridge’s text analytics and sentiment analysis functionality, check out The Truth About Text Analytics and Sentiment Analysis.