Growing class sizes and the rapid advancement of information technology have prompted institutions to implement blended learning in their academic offering. There are few reports in the literature on a holistic assessment of the effectiveness of the instructional design of blended learning courses for large classes. We defined effectiveness from a pedagogical and outcomes perspective and interrogated the design of a large first-year biology course with the aim of optimising the blend to benefit all students, but specifically the at-risk and 'murky middle' (MM) students. This was achieved by investigating patterns of student engagement with learning opportunities over time and its relationship with performance. The results show that participation in face-to-face tutorial classes was the stronger predictor of student success followed by online formative assessments. Students did not engage meaningfully with optional activities, such as virtual classrooms and pre-reading for lectures, and their engagement with compulsory learning opportunities declined during the semester, with the MM slipping towards those at risk of failing. We conclude that students need carefully chosen compulsory activities that are complementary in design and purpose, while optional activities are best suited for enrichment purposes.