This chapter describes and empirically illustrates an approach to analyzing categorial phenomena in talk-in-interaction, grounded in the distinctive conversation analytic practice of building and analyzing collections. We begin by outlining a core set of observations made by Harvey Sacks in implementing a shift from conventional social scientific treatments of categories (e.g., gender, race, sexuality, age) as analysts’ resources, to instead examining them as members’ resources. That is, instead of using categories to study the social world, Sacks’ approach introduced resources for seeing how participants in social interactions use and self-administer categories. We then present an analysis of a collection of openings of interactions from ordinary conversational and institutional settings, considering some ways in which participants explicitly and tacitly use and manage categories in the initial moments of these interactions. Using this analysis as an exemplar, we address a set of challenges and critiques associated with conversation analytic research on categories. We thereby describe how CA can provide an empirically rigorous means of examining the ‘mutually constitutive’ relationship between categories and other ‘generic’ interactional structures and practices – and thus for analyzing the situated (re)production of categories, from the most mundane to those most strongly associated with distributions of power and privilege.