How can I analyse open-text customer feedback that has no score?
Treat each comment as a statement about a need. Code it to one need and one driver, judge from the words whether the need was met, and note how strongly the customer feels. Then net the positive experiences against the negative ones for each need. This is not sentiment analysis: it tells you which needs are failing.
The Customer Needs Solution (CNS) normally asks customers to score each statement for Performance and Emotion. Where only unscored text exists, such as social posts, CNS sets out an alternative: Social NetCNS, the percentage of positive experiences minus the percentage of negative ones, per need and overall.
A throwaway mention counts as Passive. Strong language, a repeated complaint or a threat to leave makes it count. Report Social NetCNS separately from survey NetCNS, never blend the two, and calibrate it now and then against a small scored survey.
AI can speed up the first pass, but a person should check the coding: without a knowledge base of your organisation's needs and drivers, AI-generated themes can vary from one run to the next.