Data sets sometimes contain more than one column. You might have some other columns than the text verbatims such as a Rating / NPS /  Date and you have columns that are kind of like demographics data, which could include Location, or Store ID.

Using a text analytics tool to understand text data seems like a natural step after doing a survey or collecting text data. You can uncover some great Topics from your text data and get a strong understanding of what the text is about. 

Well if you kept your other columns in the analysis you can get an even deeper insight with “who” said what. We can call these columns of data, metadata columns, that give extra information about the text column, but in essence you can think of them as demographics data. For example, if your customer survey included location and years they were a customer then you could figure out the Topics, and then the breakdown of who said each Topic. Such as, the relative importance of Topics for long time customers versus newer customers or the different locations and which prefers to speak about which Topics. 

So after analyzing the text responses, you can create action items to address the biggest worries for different customer types based on the metadata you had in the survey. This will allow you to make the largest improvements to different segments and to improve your total customer satisfaction. In the long run, you can begin to segment your customers and allow for custom accommodations based on the segment. Text analytics is understanding your customers, which means both an increased satisfaction and profitability.