Community environmental management (CEM) involves the facilitation of community partnerships, local dialogue, consultation and participative decision-making. This is increasingly seen as a solution to some of the more complex environmental issues faced by regulatory authorities. Anecdotal evidence suggests that CEM programmes have much potential, but the evaluation of them is problematic, and there is a need for more robust evidence of their effectiveness. This paper reports on the development of a new CEM evaluation approach (inspired by soft systems methodology, developmental work research and systemic intervention), which was trialled with a New Zealand regional council. The approach shows promise in addressing common evaluation bottlenecks and helping stakeholders to develop causal narratives that more fully account for the complex relationship between community participation and environmental outcomes. However, while the local participants in the CEM initiative acted on the evaluation findings, they hoped that it would stimulate wider organisational change, and this did not happen. Project reflections, informed by institutional theory, reveal that the logics of 'participation' and 'community' implicit in the findings were appropriate for local participants, but non-participating regional council stakeholders read the findings with different logics, and therefore the evaluation failed to communicate the necessity for wider change. The reflections highlight a previously unrecognised evaluation bottleneck. While the CEM evaluation methodology has the potential to be adapted for other contexts, if wider organisational change is required, care must be taken to anticipate the different institutional logics of stakeholders who might be unfamiliar with, or even hostile to, CEM. (c) 2020 Elsevier B.V. All rights reserved.