Breast cancer (BC) has been classified into four intrinsic subtypes through seminal studies employing gene expression profiling analysis of primary tumours, namely the luminal A and B subtypes, the human epidermal growth factor receptor 2-like subtype and the basal-like subtype. More recently, the emergence of high-throughput genomic sequencing techniques, such as next-generation or massive parallel sequencing has expanded our understanding of the complex genomic landscapes of BC, with marked intertumour heterogeneity seen among different patients. In addition, increasing evidence indicates intratumour heterogeneity, with molecular differences observed within one patient, both spatially and longitudinally. These phenomena have an impact on the clinical development of molecularly targeted agents, with the classical paradigm of population-based clinical trials being no longer efficient. In the era of genomically driven oncology, three complementary tools can accelerate the clinical development of targeted agents for advanced BC as follows: (i) the implementation of molecular profiling of metastatic tumour lesions, as exemplified by the AURORA (Aiming to Understand the Molecular Aberrations in Metastatic Breast Cancer) programme; (ii) serial assessments of circulating tumour DNA, allowing a more thorough molecular interrogation of metastatic tumour burden; and (iii) new innovative clinical trial designs able to address the challenges of the increasing molecular fragmentation of BC.