This article analyses some typical consequences of a specific research policy on experimental research in biology. The policy is conducted by a national funding agency--the Swiss National Science Foundation--through a particular programme, the "National Centres of Competence in Research" which is designed to promote both "scientific excellence" "and" "managerial professionalism". To study the possible tension between the two objectives, as a practical matter for researchers, the proposed analysis focuses on the interaction between two laboratory scientists and the administrators of a genomic platform. Access to the instruments of this platform is granted through a preliminary interview with those in charge of the platform. During that interview, researchers are required to explain why they want to use the platform services and what their expectations concerning their envisaged activities are. A tape-recorded interview is analysed in order to describe how turns at talking by the various parties, as well as the formulation of the problems encountered by a researcher, prove category-bound. The first part of the meeting ("problem exposition") is structured by the categorical device "generalist researcher vs. specialist researcher", whilst the second part ("problem solving") is organised by the categorical device "manager vs. user of the platform". The "scientific" problem becomes a "technical" one and the choice of technique is partly based on financial reasons. The situation shows how managerial injunctions of research policy are not without practical consequences for research activities "in situ".