A variety of public, governmental initiatives inform citizens about what is considered ‘healthy’, ‘climate friendly’, and in general ‘sustainable’ food. The ambition is to influence individual food choices. However, research also suggests that, rather than public authorities, people are more likely to influence each other’s behaviour. Yet, the degree of influence seems to depend on how people are connected and how similar they are to each other. To better understand the various dynamics occurring within social networks, researchers have experimented with controlled networks, manipulating factors like the number or kinds of ties. This increases the chances of identifying causality, yet also affects the external validity of the results since the design is very artificial. The paper outlines a 4-month randomized controlled trial that tries to minimise artificiality by studying real-life behaviour, while including controlled treatments. Specifically, using a mobile application, we will study if and how real-life, self-reported intakes of plant- and animal-based foods diffuse under certain social network conditions. The participants are randomly assigned to distinct networks, each exposed to different signals. One aim is to compare whether behavioural change spreads more (if at all) in groups exposed to (1) different social network conditions (interaction possibilities by sharing pictures, likings, and comments) and (2) different source-types of information about food (fact-based versus social group-based). By investigating the roles of both social diffusion processes and certain structural conditions provided by public actors for sustainable consumption, the results contribute to discussions of how and by whom effective and efficient measures could be implemented to ease and foster sustainable consumption.