The widely accepted assumption that the development of more efficient single molecule magnets must involve ever higher total spin values has so far driven synthetic efforts towards molecular clusters of increasing nuclearity. In the present paper it is suggested that it might be worthwhile to reconsider this approach. There is evidence from theory and experiment to suggest that the race for multinuclear complexes with higher total spin might not necessarily be fruitful as a strategy for maximizing the magnetic relaxation barrier. Instead, we propose that more effort should be directed in understanding the parameters involved in maximizing the anisotropy of small, perhaps even mononuclear, molecules. Using multi-reference ab initio calculations we demonstrate the theory that can be applied and the principles of the computational approach for representative mononuclear complexes. Such small units may subsequently be employed as building blocks for the controlled assembly of larger and maximally anisotropic single molecule magnets.