1. The vast majority of species interactions are seasonally structured and depend on species’ relative phenologies. However, differences in the phenologies of species naturally vary across years and are altered by ongoing climate change around the world. 2. By combining experiments that shifted the relative hatching of two competing tadpole species across a productivity gradient with simulations of inter-annual variation in arrival times I tested how phenological variation across years can alter the strength and outcome of interspecific competition. 3. Shifting the relative timing of hatching (phenology) of a species fundamentally altered interspecific competition, and the effect of shifting the timing on competition was highly non-linear for most demographic rates. Furthemore, this relationship varied with productivity of the system. As a consequence, (i) shifts in relative timing of phenologies had small or large effects depending on the average natural timing of interactions, and (ii) changes in the inter-annual variation in onset of interaction alone can alter species interactions in simulations even when mean phenologies (timing) remain unchanged across years. 4. Traditionally, studies on phenologies focus on directional shifts in the mean of phenologies, but our results suggest that we also need to consider inter-annual variation in phenologies of interacting species to predict dynamics of natural communities and how they will be modified by climate change.