With the growing centrality of digital recording technologies to conversation analysis (CA) research, an emerging array of publications has begun to provide useful methodological insights on how to capture multimodal and temporal complexities of social interaction in video footages. By and large, however, they have been written as general guidelines, without much regard to specificities of institutional settings. To address this gap, this article discusses specific considerations for the production of audio-visual data for CA research on classroom interaction. We argue that a set of heuristic considerations is needed to prevent researchers from overlooking details that participants orient to as constitutive of their institutional activities. To this end, we offer a brief overview of common characteristics of classroom interaction. These include its multiple spatial arrangements within and across lessons and pedagogical projects which are accomplished through local actions and action sequences. Building on these characteristics, we provide a set of guiding questions to facilitate pre- and online decision-making processes that are undertaken by individual researchers in the data collection phase. Yielding unprecedented opportunities for research, careful and disciplined attention to the production of video data is indispensable as we continue to study and theorise classroom interaction in diverse contexts.