The binding energies (BEs) of alkali metal monocations and alkaline-earth metal dications to a series of small oxygen and nitrogen bases have been evaluated by means of CCSD(T) calculations on B3-LYP optimized structures. These calculations were carried out both using all-electron basis sets, and additionally using an effective-core potential (ECP) to describe the inner electrons of the metal. A theoretical model aiming at analyzing the effects on the binding energy trends of electrostatic, polarization, and covalent contributions, as well as geometry distortion, was employed. From this analysis, we conclude that although the neutral-ion interaction energy for alkali and alkaline-earth metal cations is dominated by electrostatic contributions, in many cases the correct basicity trends are only attained once polarization effects are also included in the model. This is indeed the case when Ca2+ and K+ are bound to ammonia and formaldehyde. Geometry distortions triggered by polarization are also necessary, in some cases, to obtain the correct basicity trends. In addition, in particular for alkaline-earth dications, the energy associated with covalent interactions sometimes dictates the basicity trend. Our observations imply that simple models based on ion-dipole interactions, that are frequently used in the literature to explain affinity trends in ion-molecule reactions, are generally not likely to be reliable.