To valorize lignin as a renewable source of aromatics, it is necessary to develop selective heterogeneous catalysts for the hydrodeoxygenation reaction of aromatic oxygenates such as anisole. Most of the metal supported catalysts tested so far exhibit a high conversion but a low selectivity towards valuable aromatic hydrocarbons, yielding mainly phenolic compounds. To gain insights into that catalytic system, we performed surface science experiments (X-ray Photoelectron Spectroscopy and Temperature Programmed Desorption)under Ultra-High Vacuum conditions (UHV). Dosing anisole on Pt(111) surprisingly gave benzene, carbon monoxide and hydrogen as the main desorbing products of decomposition.With the help of Density Functional Theory (DFT) we successfully explain the unexpected selectivity. In the present work we show in particular that phenoxy PhO stands as a key intermediate. Although the UHV conditions do not allow the hydrogenation of phenoxy into phenol, i.e. the catalytic product, they reveal the key role of both hydrogen and carbonaceous species. Under UHV conditions, anisole gets extensively dehydrogenated: it results in the formation of carbonaceous fragments, which can actually perform the deoxygenation of phenoxy into benzene, but also, more importantly, coke. This detailed study opens the door to a rational design of hydrodeoxygenation catalysts based on supported metals.