Despite tangible progress in recent years, substantial therapeutic challenges remain in multiple myeloma (MM), particularly for patients at high risk for early relapse or death and for those with advanced multi-drug resistant disease and refractoriness to currently available combination regimens. Addressing these challenges requires identification of novel classes of anti-MM agents, their incorporation into safe and more effective combination regimens, and development of efficient algorithms to select the most appropriate therapeutic options for the clinical and molecular features of individual patients at a given time during their disease. Ideally, these goals can be facilitated by preclinical identification of the "driver" molecular lesions on which different myeloma subtypes exquisitely depend, and by informative preclinical models simulating the clinical setting(s) in which trials will be conducted. Large prospective studies of patients treated uniformly with contemporary clinical regimens are essential, but there is also a major need for flexibility in studying new regimens in the future. Long-term patient follow-up and integrated annotation of clinical (safety and efficacy) and correlative (molecular, biochemical, etc) data are also critical. Novel molecular profiling techniques will likely identify more clinically and biologically discrete subsets of patients with recurrent, even if infrequent, lesions. This molecular heterogeneity, combined with the increasing numbers of candidate therapeutic targets and respective investigational agents, may pose formidable challenges for the development and implementation of personalized medicine in MM. This review discusses these challenges, as well as potential strategies to address them, with the aim of making significant improvement in the clinical outcome of patients with MM.