Theory suggests that the spatial context within which species interactions occur will have major implications for the outcome of competition and ultimately, coexistence, but empirical tests are rare. This is surprising given that individuals of species in real communities are typically distributed nonrandomly in space. Nonrandom spatial arrangement has the potential to modify the relative strength of intra- and interspecific competition by changing the ratio of conspecific to heterospecific competitive encounters, particularly among sessile species where interactions among individuals occur on local scales. Here we test the influence of aggregated and random spatial arrangements on population trajectories of competing species in benthic, marine, sessile-invertebrate assemblages. We show that the spatial arrangement of competing species in simple assemblages has a strong effect on species performance: when conspecifics are aggregated, strong competitors perform poorly and weaker competitors perform better. The effect of specific spatial arrangements depends on species identity but is also strongly context dependent. When there are large differences in species competitive ability, aggregated spatial arrangements can slow competitive exclusion, and so nonrandom spatial arrangement can work synergistically with other trade-off based mechanisms to facilitate coexistence.