Myeloma is a complex disease, characterized by a wide heterogeneity in clinical presentation, evolution, and molecular portraits. The successive use of cytogenetics, molecular cytogenetics, expression genomics, copy number genomics, and, more recently, deep sequencing, has shown that this heterogeneity can be used to identify markers usable for not only prognostication but also therapeutic choice and, ultimately, discovery of druggable targets. The use of some of these techniques is now mandatory for the management of patients. Although risk-adapted therapy is not yet a routine practice in myeloma, these molecular changes are essential for the definition of the prognosis.