The photo-induced reactivity of compounds at the surface of photocatalytic materials is used to maintain the cleanliness of the surface of glass, concretes and paints. A standard method to quantify the photocatalytic self-cleaning (SC) properties of non-transparent materials was recently published. It is based on the covering of the sample surface with a defined amount of dye and on the evaluation of the reflectance spectra of the coloured surface under irradiation. The calibration of the spectral changes allowed the quantification of the surface residual dye and the evaluation of the self-cleaning kinetics. The method was tested on seven white and coloured photocatalytic materials using methylene blue (MB), rhodamine B (RhB) and metanil yellow (MY). The main by-products of the MB photocatalytic degradation at the solid/solid interface were identified, showing that MB degradation in solution follows a path quite different from that at the solid/solid interface. Also MY showed a different order of photoreactivity. Furthermore, experiments at the solid/solid interface are more trustworthy than tests in solution for evaluating the self-cleaning ability. The differences of the photocatalytic phenomena at the solid/solid interface in comparison with the most studied photoactivated processes at the solid/liquid interface are outlined. Furthermore, photocatalytic materials showed selectivity toward some specific dyes. This encourages the use of more than one dye for the evaluation of the self-cleaning ability of a photocatalytic material.