Interest in host-associated microbiomes has skyrocketed recently, yet our ability to explain microbiome variation has remained stubbornly low. Considering scales of interaction beyond the level of the individual host could lead to new insights. Metacommunity theory has many of the tools necessary for modeling multiscale processes and has been successfully applied to host microbiomes. However, the biotic nature of the host requires an expansion of theory to incorporate feedback between the habitat patch (host) and their local (microbial) community. This feedback can have unexpected effects, is predicted to be common, and can arise through a variety of mechanisms, including developmental, ecological, and evolutionary processes. We propose a new way forward for both metacommunity theory and host microbiome research that incorporates this feedback.