Classification of lithic artifacts’ raw materials based on macroscopic attributes (e.g., color, luster, texture) has been used to pull apart knapping episodes in palimpsest assemblages by attempting to identify artifacts produced through the reduction of an individual nodule. These classes are termed “raw material units” (RMUs) in the Old World and “minimum analytical nodules” in the New World. RMUs are most readily defined for lithic artifacts in areas with distinctive cherts and other siliceous raw materials, allowing pieces from different nodules to be recognized visually. Opportunities to apply RMUs, however, are strongly limited at sites where lithic material visual diversity is low. The magnetic properties of obsidian, which result from the presence of microscopic iron oxide mineral grains, vary spatially throughout a flow. Consequently, obsidian from different portions of a source (i.e., different outcrops or quarries) can vary in magnetic properties. This raises the possibility that magnetic-based RMUs (mRMUs) for obsidian artifacts could be effective to distinguish individual scatters from multiple production episodes and offer insights into spatial patterning within a site or specific occupation periods. First, we assess the potential of mRMUs using obsidian pebbles from Gutansar volcano in Armenia. Second, we evaluate the validity of this approach based on a double-blind test involving an experimental assemblage of Gutansar obsidian flakes. Cluster analysis can successfully discern flakes from obsidian specimens containing high concentrations of iron oxides. Obsidian with more magnetic material has more opportunities for that material to vary in unique ways (e.g., grain size, morphology, physical arrangement). Finally, we apply the mRMU approach to obsidian artifacts from the Middle Palaeolithic site of Lusakert Cave 1 in Armenia and compare the results to traditional RMU studies at contemporaneous sites in Europe. In particular, we seek – but do not find – differences between retouch flakes (which reflect rejuvenation of tools) and the other small debris (which reflect other reduction activities). This result likely reflects the local landscape, specifically the abundance of obsidian and, thus, little pressure to curate and retouch tools. As this approach is applied to additional sites, such findings will play a central role in regional assessments about the nature and timing of the Middle to Upper Palaeolithic “transition” and the relationship, or lack thereof, between technological behaviors and presumed population dynamics.