Surface-enhanced Raman Scattering (SERS) is spectroscopic technique with strong potential for basic and applied science due to the large enhancement of the Raman signal, typically from 10^6 up to 10^{14} in the most favorable situation. This amplification is the result of several contributions, being the charge-transfer (CT) mechanism the most controversial of it due to the intrinsic properties of CT states. Specifically, the vanishing electric transition dipole moment of CT states implies that this mechanism must be active by intensity borrowing from the very bright plasmonic transitions on the metal (PL), therefore, excited state coupling must be accounted for a proper estimation of the enhancement factors. In this contribution, a survey of different computational approaches is discussed, emphasizing on the strengths and limitations of each of them. Two adiabatic methods based on simulating direct transitions on the CT states or on mixed PL-CT states are analyzed, including interstate coupling as Herzberg-Teller approximation. In addition, a full non-adiabatic approach based on diabatization procedures and nuclear wavepacket propagations with a Linear Vibronic Coupling model is also presented.