In this article we draw on lessons for collaborative lesson research (CLR) from its effective functioning as part of the design of a large-scale intervention, which aimed to improve student attainment in mathematics. The context of the research is that of post-16 students in England retaking the national examination for 16 year-olds because they failed to attain what is considered a good grade at first sitting. Improvement rates for these students is notoriously low. However, our large scale randomised controlled trial that allowed us to investigate the efficacy of CLR found that this leads to improvements in student outcomes, above and beyond teaching with access to the same accompanying professional development programme and resources. Here we reflect on what we might learn for CLR more generally.