Skip to main content

Try Qualtrics for free

Free Account

Understanding text analytics: Your definitive guide to success

13 min read
What is text analytics, and how does it differ from sentiment analysis and text mining? Read on to learn how to carry out and use text analytics for improved customer experiences.


What is text analytics?

Text analytics is the process of taking unstructured text data and analyzing it, creating pertinent business insights from the relevant information that you find. This process can be automated, powered by machine learning, artificial intelligence, and natural language processing (NLP).

A good text analytics platform enables users to understand copious amounts of text data without losing context, pull out complex concepts and discern the truth from ambiguous data in an unbiased way.

Social listening in action

With digital channels becoming the preferred method for customer communication, having a way to analyze a significant volume of textual data can help you to get deeper insights more quickly and efficiently.

See how Qualtrics and Clarabridge work together for deeper text analytics

What is text analytics used for?

4.66 billion people are using the internet alone, with 4.2 billion using social media.

Often, their interactions take the form of text. Think, social media posts, third-party review sites, emails, blog copy, and more – these can all be scoured for insights using text analytics.

You can also use text analytics for customer feedback analysis on customer survey responses and other solicited information you use to understand your audience’s point of view.

Rather than relying on humans to source, collect and analyze text data for patterns, text analytics can identify tangible insights for brands to utilize, making it one of the more efficient research methods for understanding your customers.

Users can leverage text analytics to support a multitude of objectives:

Business planning

By understanding the customer experience through data insights, you’re in a far stronger position to improve pain points in the customer journey and ultimately improve customer satisfaction. Rather than guessing at what your customers think and how they feel about your brand, you can better target your efforts to where it matters most.

Analyzing reviews on websites

Brand image

Natural language processing can help you to understand how customers feel about your brand. Customer perception can be difficult to pinpoint, so analyzing large-scale textual data can provide a more holistic view of the customer experience. Suffering from negative reviews or unkind social media posts? You can find out exactly what people are saying and flag issues as they arise.

Omnichannel Analysis

Instead of asking your team to wade through large quantities of data and manually integrate findings from across sources, you can use a text analytics platform to gather, analyze and understand large amounts of data in a comprehensive and consistent way. By using text analytics and natural language processing, your team can focus on what matters – improving customer experiences instead of data gathering and analysis.

Omnichannel analysis

Market Research

You can also use text analytics techniques to develop an in-depth understanding of your audience segments, finding details on their interests, opinions, purchasing habits, and more. This approach helps you to build more detailed buyer personas and target your customers more effectively.

Granular insights that drive action

Your customers are human and will likely give you feedback that is nuanced and occasionally difficult to categorize. A sophisticated tool analyzes ambiguous statements and detects trends across customer feedback data can offer a better understanding of customer feedback and inform next steps.

Learn how to improve your customer experience with text analytics from Qualtrics

What’s the difference between text analytics, text mining, and sentiment analysis?

Text mining vs text analytics

Text mining is the simple process behind analytics. Your textual data is analyzed by applying text analysis algorithms, producing quantitative results such as what subject is being talked about most, and where it is being talked about.

The insights portion is where the text analytics comes in. What does this information mean in real terms for your business? Analytics provides you with the qualitative and quantitative results you need to make decisions.

Read our full guide to the text analysis

Text Analytics vs. Sentiment Analysis

Text analytics interprets the data within customer feedback, finding categories, frequencies, and regular expressions in written text. Sentiment analysis helps users determine whether an expression is favorable, unfavorable, or neutral and to what degree.

Text analytics vs sentiment analysis: full breakdown

Text analytics and sentiment analysis are often positioned as interchangeable, even if they’re not.  Though both methods of analysis allow users to discern meaning when examining customer data as part of a well-executed customer experience management program, text analytics is a broader term while sentiment analysis refers to a specific type of measurement.

The key points of differentiation are:

Each unveils a distinct content type

Text analytics is a mechanism for deeper analysis of customer feedback. Using text analytics, users can dive into the data, revealing what topics are being written about most. It unveils what’s trending, which ideas are often related, and pinpoint who is highlighting which topic the most. On the other hand, sentiment analysis is a type of text analysis that specifically evaluates feedback to determine whether customers feel positively or negatively about a topic.  allows you to see if the topics are being portrayed in a positive light or a negative one. Sentiment analysis can also be applied to data that’s not in written form, such as audio or video clips or even images.

Both are able to flag issues, but do so in different ways

Text analytics can notify you when a new topic appears or starts to trend when you’re examining data. If a word with a negative connotation abruptly begins to trend in a certain area, users can quickly put their team to work resolving whatever issue is being flagged.

Sentiment analysis helps you to find potential risks in the data you’re collecting over time. For example, if you previously scored well on sentiment but the data now appears to be trending more negatively, you can investigate why this might be occurring.

Each type of analysis functions differently

Qualtrics, powered by Clarabridge, uses natural language processing algorithms to examine textual data. This allows users to perform text analytics in the same way the human brain might process the same information. Using proprietary algorithms, the XM platform can identify parts of speech, understand what’s being said and what ideas are being expressed, and then find meaning – all while checking for mistakes and self-correcting. It can offer users an expansive view across all your data, or provide analysis on a single data point.

Meanwhile, sentiment analysis can be used to examine the meaning of words or phrases and score them on a scale from positive to negative. We use an 11-point sentiment scale, which provides a sophisticated view of the customer experience.

How to perform text analytics

Text analytics is a complex process, but it is one that is made easier by choosing one of the several text analytics tools available to do the hard work for you.

However, the basics of text analytics are as follows:

Gather your data

You should gather as much data as possible to analyze using your text analytics tools. Structured data and unstructured data in multiple languages and across varying platforms should be collected to provide you with the broadest possible understanding of the big picture. It helps if all of this data can be stored in a single secure platform, so teams can access insights quicker.

Your text analytics methods will depend on what tool you use, but your text analytics process should include the following:

  • Language identification: What language is being used in this text data?
  • Part of speech tagging: Your deep learning algorithms will have been trained to tag specific words to parts of speech, such as nouns, verbs, and adjectives.
  • Sentiment analysis: It can then identify clauses or phrases within each sentence and link it to sentiment.
  • Named entity recognition/entity recognition: Using natural language processing, the tool can identify names within the text and classify them according to categories that will have been previously defined, such as person, organization, location, or monetary value.
  • Event extraction: This is where your tool gathers information about a periodical interaction, finding details on what happened and when.
  • Text classification: Using specific parameters, your text analytics tool can assign relevant tags to data, helping you to sort more easily through text.
  • Text categorization: Your text analytics tool is constantly learning, and after using training data, it will be able to categorize data in future using what it already knows.
  • Term frequency: How often does a term come up in the data you’re analyzing?
  • Topic modeling: This is where your tool finds groups of items and identifies them as a cluster or topic.

Act on insights

The data you extract will likely provide you with clear directives as to what you should be doing next. Perhaps the term “expensive” keeps coming up, or the text analytics software detects the topic of your competitor’s brand. Understanding the sentiment around these insights and taking direct action – lowering prices, or making sure you’re more attractive than your competitor – is how you can use text analytics to provide a more effective and enticing customer experience.

Examples of text analytics

Here are a couple of examples of how the Qualtrics text analytics tool can help you to develop useful business insights.

A sophisticated understanding of sentiment

By identifying sentiment in written text by looking at clauses – the distinct parts of the sentence – you’re able to understand precisely what a customer feels positively or negatively about.

For example, pretend that a customer has written the following comment in the open text section of a customer survey feedback:

“The customer service representative helped me solve my problem, but I was on hold for a disappointingly long time which ruined the experience.”

Looking at this through a sentiment analysis lens, the first clause – “The customer service representative helped me solve my problem” is positive, but the second “I was on hold for a disappointingly long time which ruined the experience” is negative.

Many text analytics platforms will score this sentence as neutral, as the positive and negative sentiments “cancel” each other out and they work to a fixed scale of positive-neutral-negative. While someone applying a three-level scoring system would have to decide whether to weigh the liking of the customer service more heavily than the difficulty of the customer service call to determine the overall sentiment,  our sentiment analysis scale allows us to break the sentence down more specifically.

This sentence expresses multiple feelings about the experience. It needs to be on a scale of positive to negative to show that although there are both positive and negative sentiments, it’s not all negative or all positive, and it’s not neutral either – there is sentiment here that shouldn’t be missed.

The phrase “helped me solve my problem” could garner a +3 score, while “ruined the experience” could get -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.

Less thorough insights, gained at scale, can lead to a lower return on investment, and more significantly, it can mean that you focus on the wrong things. As you’re likely looking at the extremes of sentiment with other tools – highly positive or highly negative – you’re not getting the nuanced picture that can help you to prioritize key areas for work.

Text analysis with Qualtrics, now powered by Clarabridge

Our XM platform leverages text analytics to conduct sentiment analysis and other data enrichment to reveal deep, granular 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.

Try the demo now 

Make sure your listening to your customers

Learn more