Research on the consequences of changes in biodiversity for ecosystem functioning is still a matter of debate, especially for microbial communities. In the special case of engineered ecosystems, e.g. anaerobic digesters, manipulated diversity experiments are scare. More efficient use of resources may result from the selection and enrichment of few individual performing microorganisms, i.e. bioaugmentation, or from the selection of a complex network with its numerous interactions. In the case of bioaugmentation, specific strains are typically isolated outside their environmental context and then mixed with a naturally assembled community. However, assembling artificial communities by mixing already naturally occurring networks may be a more successful strategy than adding individuals. Following this idea, we designed a common garden experiment where we expose mixtures of different source communities to standardized environmental conditions. The source communities come from functionally redundant anaerobic digesters. Research on the consequences of changes in biodiversity for ecosystem functioning is still a matter of debate, especially for microbial communities. In the special case of engineered ecosystems, e.g. anaerobic digesters, manipulated diversity experiments are scare. More efficient use of resources may result from the selection and enrichment of few individual performing microorganisms, i.e. bioaugmentation, or from the selection of a complex network with its numerous interactions. In the case of bioaugmentation, specific strains are typically isolated outside their environmental context and then mixed with a naturally assembled community. However, assembling artificial communities by mixing already naturally occurring networks may be a more successful strategy than adding individuals. Following this idea, we designed a common garden experiment where we expose mixtures of different source communities to standardized environmental conditions. The source communities come from functionally redundant anaerobic digesters. The selection occurred during 3 months in continuous cultivation mode. At the end of the selection period, the digesters differed in community structure and therefore likely had their own set of well-established interactions. We tested how the initial diversity of the mixtures of source communities affects ecosystem performance. The performance was estimated by the production of biogas from a standardized complex organic substrate. MiSeq inventories of partial 16S rRNA gene amplicons of Bacteria and Archaea were obtained for every source community at the day of mixing. Aliquots from the source communities were anaerobically sampled and mixed with an increasing complexity of 1, 2, 4, 8, 12 or all available communities, maintaining the same amount of total biomass per mixture. The batch production of biogas was monitored in altogether 164 mixtures, including replicates. Performance increased with the number of mixed ecosystems. The increase of performance was also related to the Shannon diversity index of initial mixtures. Likewise, we observed transgressive overyielding: more biogas was produced by the mixtures compared to the most productive source community when tested alone. The transgressive overyielding increased up to 26% with the initial diversity and with the number of source communities. Increasing diversity improves performance, but transgressive overyielding suggests likewise complementarity effects that go beyond a simple dominance of the most productive microbial community (a sampling effect). From an application perspective, the careful choice of the inoculum may be a lever for improving the final performance of anaerobic digesters. Mixing as many sludge as possible during the start-up may trigger high biomass conversion efficiencies.