The category of community health worker applied within the context of health intervention trials has been promoted as a cost-effective approach to meeting study objectives across large populations, relying on the promotion of the concept of ‘com-munity belonging’ to encourage altruistic volunteerism from community members to promote health. This community-based category of individuals is recruited to facilitate externally driven priorities defined by large research teams, outside of the target research environment. An externally defined intervention is then ‘brought to’ the community through locally recruited community volunteers who form a bridge between the researchers and participants. The specific role of these workers is context-driven and responsive to the needs of the intervention. This paper is based on the findings from an annual evaluation of community health worker performance employed as community counsellors to deliver semi-supervised HIV self-testing (HIVST) at community level of a large HIV/TB intervention trial conducted in urban Blantyre, Malawi. A performance evaluation was conducted to appraise individual service delivery and assess achievements in meeting pre-defined targets for uptake of HIVST with the aim of improving overall uptake of HIVST. Through an empirical ‘evaluation of the evaluation’ this paper critically reflects on the position of the community volunteer through the analytical lens of ‘failure’, exploring the tensions in communication and interpretation of intervention delivery between researchers and community volunteers and the differing perspectives on defining failure. It is concluded that community interventions should be developed in collaboration with the population and that information guiding success should be clearly defined.