Summary Selectivity on metal oxide catalysts is ultimately determined by complex intermolecular and surface-adsorbate interactions. Competing reaction channels are facilitated or hindered by the coordination geometry around metal cations, the ease of reduction of the surface, and the resulting stabilization of surface intermediates. The decomposition of relatively simple organic molecules like methanol and formic acid can be surprisingly complex, but attention to a few concepts may help to understand the reaction processes:o 1) The population of surface defects and coordination vacancies drives alkoxide and carboxylate formation and decomposition. When cations have at least two coordination vacancies, bimolecular reactions are possible (e.g., acetone from acetate ions, dimethylether from methoxy groups). 2) The ease of reduction of the metal oxide is often coupled to the product selectivity due to the involvement of oxygen vacancies in the reaction. For example, MgO(100) exclusively dehydrates formic acid to form CO, while the more reducible ZnO(0001)-Zn surface readily produces CO2 by dehydrogenation of formate. 3) Reduced cation oxidation states can drive reduction reactions of oxygenated adsorbates. Sputter-reduced surfaces of TiO2(001) readily converted formic acid into formaldehyde and methanol into methane and CO. 4) Interactions with other molecules can affect the ultimate reaction selectivity of the process. For example, the catalytic dehydration of formic acid on TiO2(001) occurred as a unimolecular process at high temperatures and low formate coverage, while a bimolecular dehydrogenation process dominated at near-saturation coverage of the titania crystal. Even though many of these principles appear to be applicable to for multifunctional carboxylates and alkoxides, it is important to recognize that more complex molecules may be more or less influenced by differences in surface conditions than others. Small, monofunctional molecules ably serve to highlight site requirements for some reactions, but there is no reason to expect that a large, multifunctional molecule will necessarily interact with a metal oxide catalyst as a mere combination of its functionalities. Consequently, it is important to characterize the adsorption and synthesis of large molecules where possible, to determine the limitations of the principles explored here, and to develop an understanding of adsorption and reaction characteristics that will lead to more selective catalysts.