The oxidation of stearic acid (SA) by ambient air, photocatalysed by a TiO₂ film, a common test for self-cleaning glass, often exhibits a rate of stearic acid removal, r_SA, which is proportional to the fraction of UV light absorbed, which suggests that the transport of both photogenerated electrons and holes from the bulk to the surface is very efficient. This electron, hole transport (EHT) model is tested to the extreme using highly absorbing, absorbance ≫ 2, mesoporous (sol-gel) and microporous (P25) TiO₂ films, ca. 0.5–1 μm thick, coated with a layer of SA. Scanning electron micrographs of these films show them to comprise closely connected clusters, which themselves comprise a network of primary photocatalytic particles. Each film produced identical [SA] vs irradiation time decay profiles when irradiated from the front or the back, even though in the latter case the UV irradiance reaching the photocatalyst particles in contact with the SA layer is at least 1000 times lower than the incident irradiance. These results suggest the transport of the photogenerated holes, and most likely electrons too, in these films is very efficient. When the SA layer is replaced with an Ag photocatalyst indicator ink, Ag paii, on a P25 TiO₂ film, and irradiated from the back, the ink changes colour quickly due to the reduction of Ag⁺ in the ink, indicating that the transport of photogenerated electrons through the film is also efficient. The same effect is demonstrated using highly absorbing, and scattering, films of rutile TiO₂, ZnO and WO₃, thereby suggesting the EHT model is a general feature of networked nanoparticle films. The limitations of the EHT model, and how it differs from the commonly cited antenna model, are discussed along with the relevance of these findings with regard to self-cleaning glass.