Text Mining Mainstream?
There are some rare exceptions like MotiveQuest which mines online blogs, and chat rooms to understand how consumers discuss brands in an “unsupervised” and “natural” environment online. Or SPSS which just released the first text mining application as part of their suite. But if we all agree that words are data points, then basic text mining capabilities should be part of any state of the art analytical marketing approach and team, not just a peripheral story outside of the mainstream.
How can we have real and meaningful applications of text mining methodologies as part of any innovative marketing science discipline? I see mainly three areas of applicability:
- Understand how all the different communication messages of a particular brand are disseminated throughout every channel. Most marketers of a global multi-channel brand don’t really have a good understanding of how they are messaging with consumers. A traditional brand audit does not fully capture all brand voices
- Understand how a competitor’s brand talks in every channel. Comparing one’s own brand voice with key competitor’s brand voice will reveal key differences of word clustering and might function as an early warning system if a competitor changes its brand position and language
- Understand any spoken consumer word about a company’s brand or phenomena in any channel. This could be either on the Web, in Call-Centers, in email communications, etc. It’s not just important what a brand says but also how consumers interact with a brand
The accelerating applicability of text mining will push us marketers to expand our horizon when we talk about understanding our consumers. Soon this will not be just a nice addition or gimmicky approach but an essential part of any consumer research.