Leader-member exchange (LMX) theory postulates that leaders develop relationships of varying quality with their followers. Still, leaders rarely form uniform relationship will all their followers. LMX differentiation (LMXD) is a term that captures this variation in LMX quality, which can potentially explain team dynamics and leadership effectiveness. In sum, LMXD can be viewed as a double-edged sword. When leaders thoughtfully relationship quality based on merit-based criteria, it can convey meaningful cues and enhance team effort and performance. However, if members perceive differentiation as favoritism, adverse effects can occur. The conflicting results that pinpoint its dual nature can be resolve by investigating leaders and follower traits, as well as contextual factors that moderate LMXD’s impact. Therefore, LMXD can be used strategically to allocate relational resources effectively and drive team effort, but in the wrong hands or adverse situations, it may undermine team cohesion, cooperation and performance. Rather than viewing it as uniformly good or bad, we should evaluate LMXD by considering leader-follower traits and broader contextual factors. This entry provides a theoretical clarification of LMXD, its historical evolution, antecedents, measurement approaches, and impact on individual and team work outcomes.