Data analysis in qualitative research is commonly understood as a
process that follows periods of data collection, usually carried out at a remove from the site(s) of interest. This process may be iterative, such as with grounded theory approaches, or accomplished in a particular time-period set aside for the purpose. With some methodological approaches however, this description does not suffice. Video-reflexive ethnography is a participatory and interventionist research method that engages directly and explicitly with participants to make sense of their own practices and contexts. Researchers begin with ethnographic methods such as observation and interviews, but they also concurrently work with participants to video practices of interest and/or participant generated accounts, and show this footage back to participants in researcher-guided reflexive sessions. With this method, practices of data creation and data analysis feed into one another not only iteratively, but also rapidly. Furthermore, researchers are often required to attend to both simultaneously whilst in the field, and to actively blur the lines between themselves and participants as data creators and analysts. In this paper, we draw from our collective experiences of using video-reflexive ethnography in 7 separate studies in healthcare settings, to describe how we accomplish analysis-in-action. We outline the multiple moments of analysis that occur through the method, and map them to the theoretical principles that guide these rapid, iterative and immersive research practices. We reflect on the challenges of accounting for messy research methods, but argue also that they allow us to respond more organically and dynamically to the complexity of the contexts and practices that we seek to understand.