Abstract Background Convertible stem designs allow for stem retention during revision from anatomical to reverse shoulder arthroplasty. In some cases conversion is not possible for example due to excessive soft tissue tensioning. In these cases a total revision is necessary. The primary aim of this Dutch registry study was to evaluate the unforeseen stem reversion percentages in revision of convertible anatomical shoulder arthroplasty to reverse shoulder arthroplasty. Methods Shoulder arthroplasties ( n = 2834) performed between 2014 and 2016 registered in the Dutch Arthroplasty Registry were selected. In 2016 94% of primary arthroplasties and 92% of revision arthroplasties were registered in the database. Arthroplasties were selected on convertibility. Mean follow-up was 2.4 years. We analysed the number of revisions for convertible and non-convertible designs. Cases with obligatory revisions as periprosthetic joint infections, stem loosening and periprosthetic fractures were excluded. Kaplan-Meier analysis was used to calculate humeral stem survival. Multivariate cox-regression analysis was used to determine risk factors for stem revision. Results The majority of procedures (respectively 90.9 and 72.1% for the convertible and non-convertible group) concerned a conversion to reverse shoulder arthroplasty ( p = .02). In the convertible group, the stem was retained in 29 out of 40 patients (72.5%). Overall implant survival was 94.5% after a mean follow-up of 2.4 years. Hemiartroplasty, fracture as primary indication, previous shoulder surgery and lower age were risk factors for revision. Conclusions Although convertible designs are gaining popularity due to their expected advantage in revision arthroplasty, surgeons should be aware that during a revision procedure in 27.5% of the patients an unforeseen stem revision is necessary.