With the increasing availability of therapeutic strategies (drugs, devices, disease management systems) and the growing complexity of health care delivery, there is an attendant need for objective evidence of the tangible benefits of different approaches to care. This is particularly true for patients with heart failure, a common, morbid, and resource-intensive disease. There are few well-proven therapies for patients with acute decompensation or for patients with normal LVEF. Comparative effectiveness research (CER) offers an important avenue for making progress in the field. However, CER, like any well-designed research program, requires the explicit articulation of clinically important outcomes to be compared. For patients with heart failure, there is a need to develop endpoint measures that capture the totality of potential benefits and risks for alternative therapeutic approaches. Ultimately, for one therapeutic approach to be considered superior to another, it must improve one of three relevant endpoints: make patients live longer, make them feel better, or save money without adversely affecting the other two goals. Importantly, these outcomes must be measured directly and surrogates should be avoided, even if such surrogates appear to be associated with clinically meaningful, patient-centered outcomes. In this review, we discuss the available CER endpoint domains from both a clinical and a statistical perspective, summarize the wide variety of endpoints used in CER studies, and suggest steps for greater standardization of endpoints across CER studies of patients with heart failure.