The goals of therapeutic and biomarker development form the foundation of clinical trial design, and change considerably from early-phase to late-phase trials. From these goals, decisions on specific clinical trial design elements, such as endpoint selection and statistical approaches, are formed. Whereas early-phase trials might focus on finding a therapeutic signal to make decisions on further development, late-phase trials focus on the confirmation of therapeutic impact by considering clinically meaningful endpoints. In this guideline from the Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) working group, we highlight issues related to, and provide recommendations for, the design of clinical trials on local therapies for CNS metastases from solid tumours. We discuss endpoint selection criteria, the analysis appropriate for early-phase and late-phase trials, the association between tumour-specific and clinically meaningful endpoints, and possible issues related to the estimation of local control in the context of competing risks. In light of these discussions, we make specific recommendations on the clinical trial design of local therapies for brain metastases.