Macroecology has traditionally relied on descriptive characterization of large-scale ecological patterns to offer narrative explanations for the origin and maintenance of those patterns. Only recently have macroecologists begun to employ models termed ‘process-based’ and ‘mechanistic’, in contrast to other areas of ecology, where such models have a longer history. Here, we define and differentiate between process-based and mechanistic features of models, and we identify and discuss important advantages of working with models possessing such features. We describe some of the risks associated with process-based and mechanistic model-centered research programs, and we propose ways to mitigate these risks. Giving process-based and mechanistic models a more central role in research programs can reinvigorate macroecology by strengthening the link between theory and data.