Compelling research over the past decade identified a fundamental role of the intestinal microbiome on human health. Compositional and functional changes of this microbial ecosystem are correlated with a variety of human pathologies. Metagenomic resolution and bioinformatic tools considerably improved, allowing even strain-level analysis. However, the search for microbial risk patterns in human cohorts is often confounded by environmental factors (eg, medication) and host status (eg, disease relapse), questioning the prognostic and therapeutic value of the currently available information. In addition to a better stratification of human phenotypes, the implementation of standardized protocols for sampling and analysis is needed to improve the reproducibility and comparability of microbiome signatures at a meaningful taxonomic resolution. At the level of mechanistic understanding, the molecular integration of pleiotropic signals coming from this complex and dynamically changing ecosystem is one of the biggest challenges in this field. The first successful attempts to apply reverse genetics based on the available metagenomic information yielded identification of small molecules and metabolites with functional relevance for microbe-host interactions. Further expansion on the isolation of bacteria from the “unculturable biomass” will help characterize microbiome signatures in model systems, finally aiming at the development of clinically relevant synthetic consortia with safe and functionally well-defined strains. In conclusion and beyond reasonable enthusiasm, the mechanistic implementation and clinical relevance of microbiome alterations on disease susceptibility is still in its infancy, but the integration of all the above-mentioned strategies will help overcome the correlation era in microbiome research and lead to a rational evaluation of clinical strategies relevant for targeted microbial intervention.