Text. It’s the gateway to incredible troves of first-person insight, but it can also be the bane of market research. The more text you have access to, the more you can learn – but how can you chew through enough text to get statistically significant results, without spending weeks and weeks on the process?
Adding text analytics to your market research toolbox can help you address this problem. Text analytics is the use of computer-assisted techniques that make sense of free (natural language) text and its associated metadata.
Here are 5 examples of ways that text analytics can enhance market research:
- Open-ended responses: There are some questions you just can’t get at with radio buttons and check boxes. That’s often when researchers turn to open-ended questions, only to end up flooded with hundreds or even thousands of blocks of text. Text analytics can help make sense of all that text by allowing you to human-code, machine-classify, and analyze responses when they number in the thousands or millions. Machine coding can speed the identification of topics, sentiment, and other trends in free text data.
- Leveraging metadata: If you’ve used profiling questionnaires or segmentation studies to place your customers into different audiences or demographics, you know the power of analyzing survey metadata. Using text analytics allows you to segment open-ended responses, aiding in the analysis of content you have coded or scored using custom machine classifiers.
- Disambiguation: Whether you’re analyzing text that comes from a survey, or text that comes from scanning social media, your efforts at coding or sentiment analysis may be confounded by overlapping word usage. If you’re looking at sentiment about smoking, for example, you’re interested in smoking cigarettes – not smoking guns, BBQ, or marijuana. Text analytics methods and tools sift out irrelevant references to keywords of interest, and help you focus in on the answers or mentions that you’re actually seeking.
- Brand recognition: Many companies face a very specific version of the disambiguation challenge: they’ve got a company or brand name that is a common English (or foreign-language) word, making it difficult to separate references to Avon Cosmetics from references to the Avon Theatre, Avon Indiana, and the TV character Avon Barksdale. Text analytics cuts through the noise of your social media monitoring so you tune into the references that actually refer to you.
- Historical perspective: One challenge with survey research is that you can only ask people about what they think, feel or do today – or expect to do in the future. Unless you’re doing longitudinal research with a community panel already – in which case you know the power of historical perspective – you may wonder what your customers used to think or say about you. Using a text analytics tool to dig through an archive of historical data, can give you the long-term perspective you need to understand your brand’s evolution.
These are just five of the most obvious use cases for text analytics as part of your market research toolbox. To offer these very capabilities, Vision Critical has acquired DiscoverText. As the founder and CEO of Texifter, the company that created DiscoverText, I’m delighted to be joining the Vision Critical fold, and to finding more ways of putting text analytics to work for you.