Transfusion medicine resembles all of medicine in that expert opinion predominates because hard data on clinical outcomes from randomized controlled trials and high quality observational data are simply unavailable. Indeed, some of the first trials evaluating important outcomes are barely two decades old. Patient blood management (PBM) depends on high quality data for assisting clinicians in making clinical decisions. In this review, we focus on several red blood cell (RBC) transfusion practices that new data suggest need reconsideration. The practices that may need revision include transfusion for iron deficiency anaemia, except in life threatening situations, toleration of anaemia as a largely benign condition and use of haemoglobin/haematocrit as primary indications for RBC transfusion, as opposed to adjuncts to clinical judgement. In addition, the long-standing notion that the minimum transfusion should be two units needs to be abandoned due to the danger to patients and a lack of clinical evidence of benefit. Finally, the difference in indications for leucoreduction versus irradiation needs to be understood by all practitioners. PBM is one of the strategies for managing anaemia and bleeding that holds great promise for patients, and transfusion is only one facet of the bundle of practices.