In the course of the present Debate series, several new outcome measures for assisted reproduction have been proposed to encourage the transfer of fewer embryos, in order to diminish the number of multiple pregnancies. The implementation of these recommendations, however, is hampered by the perception that safety and efficacy are communicating vessels: it is presumed that by decreasing the number of embryos transferred, pregnancy rates will decrease as well. Data from national and international registries, however, do not confirm the assumption of the communicating vessels: pregnancy rates tend to be low in countries in which many embryos are transferred, and the highest pregnancy rates occur where the number of embryos per transfer is low. Only top-level clinics (where treatment efficacy is guaranteed) are able to decrease the number of embryos transferred without compromising their pregnancy rate, and to vouch for safety in this way. Elective single embryo transfer (eSET) can never be mandatory in all patients, but the percentage of eSETs performed by a particular assisted reproduction treatment centre does reflect its quality: the ultimate outcome measure of efficacy ánd safety. Therefore, the eSET rate is the most relevant qualifier of performance in assisted reproduction.