In this paper, we show why we think the notion of instrumental ethnography should be revived (compared to Steve Woolgar's 1982 use of the term). We see instrumental ethnography as a particular form of ethnography that recognizes ethnographic stories as agential through their capacity to work as generalizations. We outline two different logics of generalizing (one-many and whole-parts) and show how these logics can be identified in a note from fieldwork that one of the authors is currently involved in. The field note displays how a project worker in a large environmental NGO engage in monitoring her partners in a development aid project; it tells about the seemingly magic actions of a database used for monitoring. We use the note for discussing why we think it is important, in a situation where ethnographic stories are bought and sold as products, to name some of the ontological commitments that go into the crafting of these stories.