The advent of precision medicine has prompted profound changes in clinical cancer research, and the rising numbers of new therapeutic agents pose challenges in terms of the most appropriate trial designs and effects on the drug-approval process. In the past 5 years, some remarkably efficacious drugs have been approved based on evidence from uncontrolled phase I trials. We challenge the view that the expected benefits from new drugs are generally sufficient to forgo a randomized trial with patients assigned to a control arm (a regimen other than the experimental treatment). Relying on efficacy results from uncontrolled clinical trials can result in expedited drug approval, but the disadvantages of this practice must be taken into account. For example, the apparent improvements in outcomes observed in an early single-arm trial of a new therapy might reflect the prognostic nature of the target, rather than a true treatment effect. Moreover, the predictive role of biomarkers cannot be definitively ascertained without randomly assigning patients to a control arm. We discuss the need for such randomization to a true control in all phases of drug development and the role of companion biomarker testing. We propose that an increased use of randomization will facilitate a seamless transition between phases of drug and/or biomarker development.