This is a study of architectural competitions as they engage with the design practices of architects within the UK and Europe. Since only one firm or one design emerges at the end, and the project programme exists prior to the submissions, there tends to be a gap between programme and practice, past and future, language and situation. It is the aim of this research to investigate what changes in our understanding of architectural practice when we acknowledge that architects work to linear programmes and submit deliverables within the set of relations that make up the competition. In conducting this research I address a gap in the social scientific understanding of architectural practice. While ethnographies of architectural studios have described the way design emerges through an interplay of humans and nonhumans, formats or structures like the competition have not yet become analytical categories in the ethnographic literature. To bridge what seems like a gap between the immaterial world of the competition and the material world of the studio, I draw from actor-network theory to view the competition as a set of relations that include objects and practices. Considering the technology of the competition, I follow five different strands of research. I identify the matters of concern that architects talk about when they talk about competitions; examine the documents involved in administering a competition; follow an atelier at an architectural school where students participate regularly in competitions; observe the Office of Metropolitan Architecture prepare a concept design; and visit an exhibition of submissions. Here I describe the ways in which competitions come together within the practice of architects. This study makes three contributions. First, the study adds to our understanding of architecture as a set of relations, rather than a stable identity. The second contribution has to do with language and practice, demonstrating that 'big' categories like 'building' nevertheless act within collectives of architects, clients, contractors and so on. A final implication is for methods. Since certain categories exist between sites, organising the activity of actors in different offices across what might be hundreds of miles, ethnographic fieldwork on architecture can become fragmented and multi-sited. The implications of the architectural competition for an ethnographic understanding of architectural practice, then, are to see more and 'bigger' collectives within the lives of architects.