A recent analysis from the PHENIX collaboration on direct photon production yield in heavy ion collisions has shown a universal, within experimental uncertainties, multiplicity scaling, in which the photon $p_T$-spectra for transverse momenta up to 2 GeV/c are scaled with charged-hadron pseudorapidity density at midrapidity raised to the power $\alpha=1.25$. This low-$p_T$ scaling suggests that the main photon sources contributing to it, could be similar across beam energies. On the other hand, particle production in large and small system collisions, including direct photons, exhibits geometrical scaling in a similar $p_T$ range. The geometric scaling follows from gluon saturation and collision geometry. In particular, the direct photon spectra obtained for large and small system collisions at different centrality classes and various beam energies depend on a specific combination of number of participants, beam energy, and saturation scale - rather than on all these three variables separately. In our presentation, we show that the multiplicity and geometric scaling laws for direct photons are interconnected and discuss physical conditions needed to relate one to another. This interrelation may help us better understand the direct photon production during the early evolution of the matter produced in heavy ion collisions.