Improvements in instructional coherence have been shown to have large impacts on student learning, yet analysis of such coherence, especially in developing countries and at a systems level, is rare. The authors use an established methodology, the Surveys of Enacted Curriculum (SEC), and apply it to a developing country context to systematically analyse and quantify the content and coherence of the primary curriculum standards, national examinations and actual teaching delivered in the classroom in Uganda and Tanzania. The authors find high levels of incoherence across all three instructional components. In Uganda, for example, only four of the fourteen topics in the English curriculum standards appear on the primary leaving exam, and two of the highest-priority topics in the standards are completely omitted from the exams. In Tanzania, only three of fourteen English topics are covered on the exam, and all are assessed at the 'memorisation' level. Rather than aligning with either the curriculum standards or exams, teachers’ classroom instruction is poorly aligned with both. Teachers tend to cover broad swathes of content and levels of cognitive demand, unrelated to the structure of either the curriculum standards or exams. An exception is Uganda mathematics, for which standards, exams and teacher instruction are all well aligned. By shedding light on alignment deficits in the two countries, these results draw attention to a policy area that has previously attracted little (if any) attention in many developing countries’ education policy reform efforts. In addition to providing empirical results for Uganda and Tanzania, this study provides a proof-of-concept for the use of the SEC methodology as a diagnostic tool in developing countries, helping education systems identify areas of instructional (in)coherence and informing efforts to improve coherence for learning.