Semiconductive metal oxides are of great importance in environmental remediation and electronics because of their ability to generate charge carriers when excited with appropriate light energy. The electronic structure, light absorption and charge transport properties of the metal oxides have made possible their applications as photocatalysts. Recently, facet-engineering by morphology control has been intensively studied as an efficient approach to further enhance their photocatalytic performance. However, various processing steps and post-treatments used during the preparation of facet-engineered particles may generate different surface active sites which may affect their photocatalysis. Moreover, many traditional techniques (PL, EPR, XPS and Raman) used for materials characterization (oxygen vacancy, hydroxyl group, cation…etc.) are not truly surface specific but the analyses range from top few layers to bulk. Accordingly, they can only provide very limited information on the chemical states of the surface active features and distributions among facets, causing difficulty to unambiguously correlate facet-dependent results with activity. As a result, this always leads to different interpretations amongst researchers during the past decades. In this article, we will review on the controversies generated among researchers, when they correlated the performance of two most popular photocatalysts, ZnO and TiO2 with their facet activities based on characterization from the traditional techniques. As there are shortcomings of these techniques in producing truly facet-dependent features, some results can be misleading and with no cross-literature comparison. This review is also focussed on the new capability of probe-molecule-assisted NMR which allows a genuine differentiation of surface active sites from various facets. This surface-fingerprint technique has been demonstrated to provide both qualitative (chemical shift) and quantitative (peak intensity) information on the concentration and distribution of truly surface features. In light of the new technique this article will revisit the facet-dependent photocatalytic properties and shed light on these issues.