The place of grammar within the teaching of writing has long been contested and successive research studies have indicated no correlation between grammar teaching and writing attainment. However, a recent study has shown a significant positive impact on writing outcomes when the grammar input is intrinsically linked to the demands of the writing being taught. The study adopted a mixed methods design with a large-scale randomised controlled trial accompanied by a qualitative dataset, which provided contextual information about how the intervention was implemented. In this paper, we will outline the pedagogical principles that underpinned the intervention, and illustrate both the theoretical grounding and practical classroom examples that exemplify the approach. We will argue that any future policy or professional development that draws on this research must take account of these pedagogical principles, rather than focusing too superficially on either the grammar or the teaching materials which exemplify them. (Contains 7 tables.)