It is becoming increasingly recognized that conventionally derived outcome measures in mental health research are problematic. This is both because of the methodology used and because a 'good' outcome is framed from the perspective of clinicians and researchers. This paper describes a methodology for developing outcome measures for use in large studies entirely from the perspective of mental health service users. It is a mixed methods model starting with a participatory and qualitative methodology and proceeding to psychometric testing. At all stages, the researchers are themselves mental health service users. In the first phase of the model, focus groups are convened comprising people who have received the treatment or service being measured. The focus groups meet twice resulting in a draft mixed-methods questionnaire devised from thematic analysis of the focus group data. This is then taken to expert panels, again comprising individuals who have received the treatment or service being evaluated for refinement. Following this, a feasibility study is conducted with N ∼ 50 participants and changes made iteratively to the questionnaire in light of feedback. The final measure is subject to psychometric testing both to ensure it is robust and to explore similarities and differences with conventionally derived measures.