A major roadblock in realizing the large-scale production of hydrogen via electrochemical water splitting is the lack of cost-effective and highly efficient catalysts for the oxygen evolution reaction (OER). In this regard, computational research has driven important developments in the understanding and the design of heterogeneous OER catalysts by establishing linear scaling relations. These relations are of paramount importance since they drastically reduce the amount of time required to traverse the vast chemical search space of potential OER materials. In this work, we interrogate 17 of the most active molecular OER catalysts known to date based on different transition metals (M= Ru, Mn, Fe, Co, Ni, and Cu), and show that they obey the linear scaling relations established for metal oxides. This demonstrates that the conventional OER descriptor established for heterogeneous systems can also be applied to rapidly screen novel molecular catalysts. However, we find that this descriptor underestimates the activity of some of the most active OER complexes as it does not consider the additional one-electron oxidation that these undergo prior to O–O bond formation. Importantly, we show that this additional step allows certain molecular catalysts to circumvent the “overpotential wall” observed for heterogeneous systems (i.e. 370 mV), leading to an enhanced performance in agreement with experimental observations. To describe the activity of such highly active catalysts, we propose a new OER descriptor that opens up the possibility of designing molecular catalysts exhibiting zero theoretical overpotential. With all this knowledge, we establish the fundamental principles for the rational design of ideal OER catalysts to advance the development of water splitting technologies.