The well-documented significance of microorganisms to the function of virtually all ecosystems has led to the assumption that more information on microbiomes will improve our ability to understand and predict system-level processes. Notably, the importance of the microbiome has become increasingly evident in the environmental sciences and in particular ecosystem ecology. However, translating the ever-increasing wealth of information on environmental microbiomes to advance ecosystem science is proving exceptionally challenging. One reason for this challenge is that correlations between microbiomes and the ecosystem processes they influence are often reported without the underlying causal mechanisms. This limits the predictive power of each correlation to the time and place at which it was identified. In this paper, we assess the assumptions and approaches currently used to establish links between environmental microbiomes and the ecosystems they influence, propose a framework to more effectively harness our understanding of microbiomes to advance ecosystem science, and identify key challenges and solutions required to apply the proposed framework. Specifically, we suggest identifying each microbial process that contributes to the ecosystem process of interest a priori. We then suggest linking information on microbial community membership through microbial community properties (such as biomass elemental ratios) to the microbial processes that drive each ecosystem process (e.g. N -mineralization). A key challenge in this framework will be identifying which microbial community properties can be determined from the constituents of the community (community aggregated traits, CATs) and which properties are unable to be predicted from a list of their constituent taxa (emergent properties, EPs). We view this directed approach as a promising pathway to advance our understanding of how microbiomes influence the systems they inhabit.