There is great interest in the use of diets to promote health via mechanisms dependent on microbial metabolism, particularly using glycans (fibre). However, both host and microbial responses to dietary glycan supplementation are variable, often poorly predictable and the underlying mechanisms for this not well understood. We postulate that a significant component of this variability arises from interactive effects of other diet components (especially protein) with microbial community assembly processes. Our aim is to identify the ecological mechanisms that constrain the host and microbiome response to dietary fibre components and elucidate design principles to improve processed food quality. Our mouse model data shows that the glycan profile of a diet is a significant driver of microbial community response, but that net outcomes for both microbes and host health are strongly interactive with protein intake and with environmental history. The same dietary fibre can be good in one diet pattern and neutral or even bad in another, but a set of rules for predicting how to optimise outcomes is emerging.