Singlet–triplet interconversions (intersystem crossing, ISC) in organic molecules are at the basis of many important processes in cutting-edge photonic applications (organic light-emitting devices, photodynamic therapy, etc.). Selection rules for these transitions are mainly governed by the spin–orbit coupling (SOC) phenomenon. Although the SOC relies on complex relativistic phenomena, theoreticians have, with time, developed increasingly sophisticated and efficient approaches to gain access to a satisfactory evaluation of its magnitude. However, recent works have highlighted the remarkable and somehow unexpected efficiency of dimers of small conjugated molecules in terms of ISC quantum yields, whose origin has not been completely investigated. In this work, we bring a coupled experimental and theoretical analysis of the origin of the unusually large ISC efficiency on a series of such dimers that differ by their nature (covalent or supramolecular). We show that considering the dynamical nature of the SOC, and especially its dependence on angular orientations between the dimer subunits sometimes overlooked in the literature, it is necessary to rationalize some counterintuitive experimental observations. This combined experimental and theoretical work paves the way for new molecular engineering rules for SOC control.