A guiding objective of phase III clinical trials should be to achieve reliable evaluation of safety and efficacy of interventions in a real-world setting, ie, as the interventions would be delivered in clinical practice. This guiding objective motivates several principles of design. The control regimen should truly reflect standard of care; in turn, one should be reluctant to use placebos that might not be inert or that might adversely impact quality of life. The choice of concomitant therapy should be sufficiently flexible as to allow for treatment options that would be widely representative of routine care. Eligibility criteria should be as inclusive as the intended labeling, specifically excluding only those patients expected to be at high risk for toxicity or have a low likelihood of benefit. Study procedures should be simplified to avoid unnecessary inconvenience to patients and investigators and to reduce trial costs. Clinical end points should be chosen that unequivocally reflect tangible benefit to patients. Experiences from the phase III trial of trastuzumab (Herceptin; Genentech, San Francisco, CA) in metastatic breast cancer, a human immunodeficiency virus/acquired immunodeficiency syndrome trial, and a trial in cardiology will be used to provide additional motivation for several of these principles. These trials provide valuable insights into the challenging issues arising in the design and conduct of clinical trials.