Statistically speaking you’ve selected a random sample out steam and took the responses that was given spontaneously and in a open ended questionnaire. What you’ve done we can translate to IDI (in depth interviews) but without any structured guideline (which is highly not recommended).
Guidelines would limit the set of responses. If you are out fishing for the most comprehensive list of input no limits should be applied.
You can then take items of interest gathered from such a wide sample and develop a more limited approach and take various customer groups into consideration.
have no statistical meaning.
That is not entirely true. Items with extremely low frequency are to be considered irrelevant.
