While government funding for preventive health has been unstable in recent times, particularly in Australia, social marketing campaigns remain at the forefront of these efforts. The knowledge produced by social marketing forms the basis of some of the most powerful discourses around physical activity. To understand these discourses requires an unravelling of the planning and design phase of a campaign which is rarely done. This paper envisages the physical activity promotion campaign as the tip of the iceberg and aims to investigate what sits beneath. In particular, we seek to follow the chain of evidence from scientific facts and epidemiological data (and what counts as such) as it is translated and transformed through preventive health experts and marketing departments to produce a campaign. We argue that commonly accepted scientific evidence on the risks of physical inactivity and in particular, its purported relationship with obesity, is complex and contradictory. Furthermore, it may also be seen as unpalatable to the broader population who are directed to change their behaviours. Therefore, what begins as epidemiological data on physical inactivity, poor nutrition, and obesity, becomes guidelines and policies about health behaviours, which are then further recontextualised into campaign marketing messages. With each move or translation, the message and its medium change to match the needs and goals of their users at each stage. As Latour illustrates, there can be no transportation without transformation. Thus, we ask three questions. What scientific evidence is cited in public health campaigns about physical activity? How does that evidence change at each stage of the planning and delivery process? And what is gained and lost once it becomes the final product of the campaign? We trace this process backwards from the messages and strategies of three campaigns: the UK’s ‘Change4Life’, and Australia’s ‘Measure Up’ and ‘Healthier.Happier.’ Data is reported from an in-depth analysis of the texts informing each of these campaigns (e.g. briefing reports, obesity statistics, public consultation). We argue that the process of promoting physical activity increasingly involves the negotiation of a complex interplay between evidence, experts, and technology. Moreover, we conclude that carefully tracing and understanding the translations entailed in these negotiations opens up spaces for considering how campaigns can be composed and enacted differently.