This case study discusses how video-recorded data make available members’ work-site practices for description and analysis. Video-recorded materials constitute what Hugh Mehan (1978) called “retrievable” data, in that these are available for repeated inspection, whereby we can subject the data to increasingly fine-grained analysis. Taking an ethnomethodological approach, our project (“Investigating Guided-Tour Interactions in Astronomical Observatories”) aims to describe the situated production of order, discourse and action—and thus the learning that goes on there—as local ensembles of competent practice and practical orientations by visitors and guides during observatory tours. Therefore, we are engaged in explicating how visitors (as lay persons) learn how (and what) to observe in a visit to the observatory, for example, how to manipulate a telescope and what to observe when peering through it. In investigating guided-tour interactions in observatories, we are able to describe the sociology of guided visits and observe how these visits are interactionally accomplished as practices, produced through various local methods invoked and used by people. To do that, however, we need to use video recordings as data, since these allow us to preserve the phenomena that happen during the guided-tours we observe.