We study how group formation and differentiated instruction affect learning outcomes. In particular, the effect of within-class heterogeneous and homogeneous group formations and differentiated instruction are evaluated on the proficiency level of eighth and ninth grade students in a financial education programme. Our paper provides evidence on both effects using two experiments involving 69 schools and 2,407 students. The identification strategy relies on a random allocation of schools to experimental conditions. The results suggest that, overall, the programme increases student financial proficiency by almost a quarter of a standard deviation. Although no main effects of group formations and differentiated instruction are found, the effect is heterogeneous as low ability students significantly benefit from homogeneous group formation and differentiated instruction.