Light has momentum. But over a century after the experimental confirmation of radiation pressure, and on the heels of Ashkin’s Nobel Prize for exploiting this effect in optical tweezers, we may still not have a detailed understanding of how it works. How, exactly, does a simple metal feel radiation pressure? Presumably, of the microscopic constituents of the metal, the free electrons experience this force first and are pushed in its direction. After all, free electrons are responsible for the conductivity leading to reflection and absorption of the light. The resulting electrical current, generated by the momentum absorbed from light, is known as the photon-drag effect. Despite its seemingly straightforward nature, which one may estimate purely on the grounds of momentum conservation, existing experiments of photon drag are rife with sign discrepancies. In this presentation, I will describe how we have solved this problem in the literature by revealing the effect’s strong dependence on surface adsorbates. This finding both impels reinterpretation of most of the published experiments and suggests a new surface characterization technique. Moreover, having resolved the sign discrepancies for pristine metal, we are forced to take seriously an unexpected finding that the intrinsic direction of the photon-drag current is opposite what one would expect for direct momentum transfer to free electrons. This result challenges the intuitive notion that radiation pressure acts on free electrons in a simple manner, raising new questions on the nature of momentum transfer in light-metal interaction.