Making use of the publicly available 1D photoionization hydrodynamics code ATES we set out to investigate the combined effects of specific planetary gravitational potential energy (φ p ≡ GM p /R p ) and stellar XUV irradiation (F XUV ) on the evaporation efficiency (η) of moderately-to-highly irradiated gaseous planets, from sub-Neptunes through hot Jupiters. We show that the (known) existence of a threshold potential above which energy-limited escape (i.e., η 1) is unattainable can be inferred analytically, by means of a balance between the ion binding energy and the volume-averaged mean excess energy. For φ p > ∼ φ thr p ≈ [7.7 − 14.4] × 10 12 erg g −1 , most of the energy absorption occurs within a region where the average kinetic energy acquired by the ions through photo-electron collisions is insufficient for escape. This causes the evaporation efficiency to plummet with increasing φ p , by up to 4 orders of magnitude below the energy-limited value. Whether or not planets with φ p < ∼ φ thr p exhibit energy-limited outflows is regulated primarily by the stellar irradiation level. Specifically, assuming photo-ionization equilibrium and setting the photo-heating rate equal to radiative losses in Lyα yields the following analytical threshold:Above this value, Lyα losses overtake adiabatic cooling and the evaporation efficiency of low-gravity planets, too, drops below the energy-limited approximation, albeit remaining largely independent of φ p . Further, we show that whereas η increases as F XUV increases for planets above φ thr p , the opposite is true for low-gravity planets (i.e., for sub-Neptunes). This behaviour can be understood by examining the relative fractional contributions of adiabatic vs. Lyα cooling as a function of atmospheric temperature. This novel framework enables a reliable, physically motivated prediction of the expected evaporation efficiency for a given planetary system; an analytical approximation of the best-fitting η is given in the Appendix.