Limited data exist on platelet transfusion during postpartum haemorrhage. We retrospectively analysed a consecutive cohort from a single centre of 347 women with moderate or severe postpartum haemorrhage, transfused according to national guidelines. Twelve (3%) women required a platelet transfusion. There were no differences between women who did and did not receive platelets with respect to age, mode of initiation of labour or mode of delivery. Women receiving a platelet transfusion had a lower median (IQR [range]) platelet count at study entry than women who did not receive platelets before haemorrhage (135 (97–175 [26–259])×109.l−1 vs 224 (186–274 [91–1006])×109.l−1), respectively), and at diagnosis of postpartum haemorrhage (median 114 (78–153 [58–238])×109.l−1 vs 193 (155–243 [78–762])×109.l−1 respectively). Six women were thrombocytopenic pre-delivery. The cause of haemorrhage that was associated with the highest rate of platelet transfusion was placental abruption, with three of 14 women being transfused. If antenatal thrombocytopenia or consumptive coagulopathy were not present, platelets were only required for haemorrhage > 5000 ml. Early formulaic platelet transfusion would have resulted in many women receiving platelets unnecessarily. Using current guidelines, the need for platelet transfusion is uncommon without antenatal thrombocytopenia, consumptive coagulopathy or haemorrhage > 5000 ml. We found no evidence to support early fixed-ratio platelet transfusion.