We present a methodological reflection that contributes to the debate on 'how to' incorporate systems thinking (ST) in education policy assessment. ST has permeated disciplines like medicine, agriculture and urban studies. Yet, its influence in education is still mild, despite claims by scholars on its potential in studying the underlying mechanisms of educational change. The ideas we discuss emerged in the context of the evaluation of a large-scale intervention in Colombia between 2018 and 2019, where we endorsed Realist Evaluation's (RE) principles to translate abstract ST considerations into a researchable evaluation design. First, we expand our discussion on the parallels between ST and RE and emphasise how the latter's tools helped apply abstract ST considerations in a practical evaluation setting. Afterwards, we document our assessment experience, accentuating our methodological decisions during its different stages: defining the system's boundaries (or components), sampling subregions of the country to visit during fieldwork, establishing data analysis protocols and conveying findings to policymakers and implementers. Based on our experience, we close by highlighting the potential of ST-based evaluation designs to engender meaningful debates among policymakers beyond narrow technical discussions of isolated intervention components to consider the interaction of multiple levels and elements of education systems.