Medulloblastoma, the most common malignant paediatric brain tumour, is currently diagnosed and stratified using a combination of clinical and demographic variables. Recent transcriptomic approaches have demonstrated that the histological entity known as medulloblastoma is comprised of multiple clinically and molecularly distinct subgroups. The current consensus is that four defined subgroups of medulloblastoma exist: WNT, SHH, Group 3, and Group 4. Each subgroup probably contains at least one additional level of hierarchy, with some evidence for multiple subtypes within each subgroup. The demographic and clinical differences between the subgroups present immediate and pressing questions to be addressed in the next round of clinical trials for patients with medulloblastoma. Many of the genetically defined targets for rational medulloblastoma therapies are unique to a given subgroup, suggesting the need for subgroup-specific trials of novel therapies. The development of practical, robust and widely accepted subgroup biomarkers that are amenable to the conditions of a prospective clinical trial is, therefore, an urgent need for the paediatric neuro-oncology community. In this Review, we discuss the clinical implications of molecular subgrouping in medulloblastoma, highlighting how these subgroups are transitioning from a research topic in the laboratory to a clinically relevant topic with important implications for patient care.