In our original editorial (Sutherland et al., 2017), we argued that co-assessment of evidence with a range of stakeholders, including local communities, would typically represent the most cost-effective way of doing conservation science. Co-assessment involves considering what works, and what does not, based on a systematic collation of global and regional datasets alongside other forms of knowledge, such as the experience of local communities. Under circumstances where there is good evidence of what is likely to work in a particular place, it does not seem efficient to co-produce new knowledge in the form of field-based scientific experiments involving local stakeholders. Where there is more limited evidence of conservation effectiveness, the issue is of sufficient societal importance, and the resources are available, then we stated it may make sense to co-produce new knowledge alongside local stakeholders.