Introduction: Increase in the number of primary shoulder arthroplasty has led to an increase in the number of revisions which presents many complex challenges and often has inferior outcomes. Materials and methods: Data was collected retrospectively, and patients were classified using Dines classification. Comprehensive case reviews were done to identify preoperative and intra-operative challenges. The primary outcome measure was Oxford shoulder score (OSS). The secondary measures were range of motion (ROM) and patient satisfaction (very satisfied, satisfied, not satisfied or worse). Results: A total of 32 patients were identified with a mean age of 67.64 years and the most common cause of revision was a combination of bone and soft tissue failure (39.3%). All patients (n=8) with hemiarthroplasty had rotator cuff deficiency while patients with resurfacing had both rotator cuff failure and bony erosion. Four patients needed a proximal humeral osteotomy and six patients needed allograft reconstruction of the glenoid for bone loss. Twentyone shoulders were revised to reverse total shoulder arthroplasty (TSA), 2 to anatomical TSA and 5 were left with cement spacer in situ. Mean duration of follow-up was 41.6 months. Mean OSS at the last follow-up was 26.88 with statistically significant improvement in ROM. There was no statistical difference in clinical outcomes (p>0.05) based on the type of primary prosthesis or cause of revision. A total of 70% patients were pain free. Patients with infection had inferior outcomes with a mean OSS of 17. Conclusion: Management of patients with failed shoulder arthroplasty is often challenging but has good clinical outcome except in infections.