Aims. Rotator cuff tendinopathy has a multifactorial origin. Rejecting the mechanistic theory has also led to abandoning operative treatment at initial presentation in the first line. Physiotherapy exercise programmes are the accepted first line treatment. The aim of this study was to assess the long-term additional benefits of subacromial decompression in the treatment of rotator cuff tendinopathy. Patients and Methods. This randomised controlled trial of 140 patients (52 men, 88 women, mean age 47.1 years; 18 to 60) with rotator cuff tendinopathy extended previous work up to a maximum of 13 years. The patients were randomised into two treatment groups: arthroscopic acromioplasty and a supervised exercise treatment and a similar supervised exercise treatment alone. Self-reported pain on a visual analogue scale (VAS) was the primary outcome measure. Secondary measures were disability, working ability, pain at night, Shoulder Disability Questionnaire score and the number of painful days during the three months preceding the final assessment. Results. A total of 90 patients (64%) returned questionnaires at a mean 12 years after randomisation. On an intention-to-treat basis, both treatment groups reached statistically significant improvement compared with the initial VAS for pain, but there was no significant difference between groups. The same was true in the secondary outcome measures. Due to group changes, the results were also analysed per protocol: operated or not. No significant differences between the groups were found. Conclusion. The natural history of rotator cuff tendinopathy probably plays a significant role in the results in the long-term. Even though the patients who underwent operative treatment had a stronger belief in recovery, which is likely to be surgical and the effect of placebo, the exercise group obtained similar results. In the future, an optimum exercise regime should be searched for, as the most clinically and cost-effective conservative treatment for rotator cuff tendinopathy. Cite this article: Bone Joint J 2017;99-B:799–805