As you’ve already seen, data is analyzed in Keatext by breaking down each sentence of long-form written feedback into statements, identifying a Topic and an Opinion for each, and then bucketing the statement into one of four sentiments: Praises, Problems, Questions or Suggestions.
You’ve probably noticed a fifth category called Mentions. This category may not be enabled for all users. If you have it, however, you are probably wondering what it’s good for.
This category consists of statements containing a Topic with no associated Opinion. For example, the statement “Some tips where to go” contains a Topic (tips) but doesn’t express an Opinion about the tips — they aren’t described as “good tips” or “juicy tips.” Single-word responses containing only a Topic are also included under Mentions.
A single word response containing only an Opinion e.g. "Great' will be never be counted as a Mention instead it will fall under the corresponding sentiment, in this case Praise.
What use is this?
Mentions are useful if you need to know how often something is mentioned in the feedback. For example, you may have a survey on hair care products where one of the questions is “What is your favourite hairspray and why?” Some respondents may ignore the “why” part of the question and just list their favourite hairspray — a one-word response. Including Mentions in your analysis allows you to see how many times that particular hairspray has been indicated by the respondents, whether or not they say why.
Another example is help desk tickets. Customers are often prompted to enter their reason for seeking help — it could be “wifi connectivity” or “billing.” They may often type in one-word responses rather than a full sentence. Using mentions in this case is a good way to easily identify which issues have the highest frequency and focus on those.
Keep in mind that if tallying frequency is no of interest to you, you can alway choose to deselect this category and conduct a complete analysis without it.