There are few independent studies reporting on the relation between interaction designers existing practices contra their efficient use of personas. Practitioners have reported that the technique have a potential for software mass-market products, and as a user-information-needs-model. Based on three different software development project characteristics this paper presents an independent study on the usefulness of personas. In the first project, it is demonstrated that personas played a surprisingly little role compared to other design influences. In the second project personas was useful as a political instrument. In the third project the balance between invested personas efforts and its usefulness was questioned. All projects demonstrate that personas functioned as an internal concept in the discussions between interaction designers and to justify design rationales 'after the fact' to other project members, as if they actually were based on the persona.