Objectives Explore how people perceive different labels for rotator cuff disease in terms of words or feelings evoked by the label and treatments they feel are needed. Setting We performed a content analysis of qualitative data collected in a six-arm, online randomised controlled experiment. Participants 1308 people with and without shoulder pain read a vignette describing a patient with rotator cuff disease and were randomised to one of six labels: subacromial impingement syndrome, rotator cuff tear, bursitis, rotator-cuff-related shoulder pain, shoulder sprain and episode of shoulder pain. Primary and secondary outcomes Participants answered two questions (free-text response) about: (1) words or feelings evoked by the label; (2) what treatments they feel are needed. Two researchers iteratively developed coding frameworks to analyse responses. Results 1308/1626 (80%) complete responses for each question were analysed. Psychological distress (21%), uncertainty (22%), serious condition (15%) and poor prognosis (9%) were most often expressed by those labelled with subacromial impingement syndrome. For those labelled with a rotator cuff tear, psychological distress (13%), serious condition (9%) and poor prognosis (8%) were relatively common, while minor issue was expressed least often compared with the other labels (5%). Treatment/investigation and surgery were common among those labelled with a rotator cuff tear (11% and 19%, respectively) and subacromial impingement syndrome (9% and 10%) compared with bursitis (7% and 5%). Conclusions Words or feelings evoked by certain labels for rotator cuff disease and perceived treatment needs may explain why some labels drive management preferences towards surgery and imaging more than others.