The ANRS-Coquelicot survey has been carried out in France for 25 years, to monitor trends in infectious diseases (HIV and hepatitis B and C) among people who use drugs. In this article, we propose to open the black box of this monitoring experience, by describing and analysing some methodological, ethical and political issues involved in this type of survey. The ANRS-Coquelicot survey has carried out on five occasions in France (from 2002 to 2023) in several cities (from 1 to 27) among people who use drugs recruited in a large diversity of services including drug treatment centres, harm reduction facilities, residential services as well as outreach teams. The survey design included a Time Location Sampling combined with the Generalized Weight Share Method and was guided by a previous anthropological survey. Over time, some methodological challenges have been addressed such as convincing harm reduction providers and drug treatment professionals to welcome this sero-prevalence survey that was first envisioned as a form social control of people who use drugs as well as an evaluation of their professional practices. The implementation of a previous socio-anthropological phase played a major role in favouring the acceptability of the survey as well as in the improvement of the quantitative methodology for choosing the most adequate methodological tools for measuring risk exposure as well as allowing participants room to promote their experiences of the survey through self-narration. Conducting sero-prevalence surveys among time is a way to produce scientific knowledge but can also inform drug policies in order to enable more favourable environments for people who use drugs. This article is an opportunity to show the richness of surveillance surveys and how they can both produce new knowledge and be the subject of innovative methodological developments, as well as enabling the evaluation and orientation of harm reduction policies over time.