Background and aims <p>Soil microbes perform essential functions for agroecosystems, but it is unclear whether and when variation in the composition of the microbial community matters for agronomic outcomes since high functional redundancy may render compositional variation irrelevant for function. We hypothesized that the complexity of organic soil amendments would determine whether microbial community composition and richness matters for efficient decomposition and ultimately crop growth.</p> Methods <p>To test this hypothesis, we conducted a greenhouse experiment in which we grew maize and rye plants in pots inoculated with soil from 11 different farm fields, each grown with one of four nutrient sources which ranged in recalcitrance to microbial decomposition.</p> Results <p>As predicted, we found that crop growth responded more positively to live vs. sterilized soil, and the microbial inocula source explained more of the variation in crop biomass when nutrients were supplied as plant litter rather than a simple mineral form. We found that the initial richness of the bacterial community correlated positively with crop growth in the presence of complex organic, but not simple mineral, nutrient sources. Differences in management practices at the source farms, however, did not explain the microbial effects on plant growth.</p> Conclusion <p>These results suggest the make-up of soil microbial communities matters for maize and rye growth but only when these crops depend on nutrients from complex organic matter.</p>