Today, companies that implement relationship and transactional surveys often rely on text analytics to synthesize and analyze customer comments. When using text analytics with survey responses, it's important to consider the nature of the open-end question that was asked.
For example, when companies ask the Net Promoter question, "How likely are you to recommend us?" a common follow-up question is, "Why did you say that?" While text analytics is valuable for categorizing the responses to this type of follow-up quesiton, you should use caution when interpreting the sentiment.
For example, someone might type "Best practice sharing" as a reason for not recommending the company. The sentiment analysis might show this response as positive (because the word "best" implies good), but when you consider the response relative to the closed-end question, it suggests "Best practice sharing" is the reason for not recommending the company and therefore should not have a positive sentiment.
Make sense?
Instead of reporting sentiment alone, consider using the response to the closed-end question, "How likely are you to recommend?" as the sentiment. You can create a graph that shows the categories from the text analytics for Promoters versus Detractors.