The narrative that ‘meat and milk are bad’ because livestock production is a major greenhouse gas emitter is widespread, promoted by international agencies, campaign groups, corporations and governments. This overarching narrative has led to generalised policy prescriptions, applicable to some western diets and to some forms of livestock production. Of course, caveats are sometimes applied, but policy and media messages tend to simplify, meaning that the vast differences between industrial and extensive livestock production are often neglected in policy and campaign messages. As a result, inappropriate policies could do great damage to livelihoods, landscapes and the life chances of people reliant on extensive livestock production, including pastoralism. Such systems involve many millions of people across rangelands covering over half the world’s land surface. Where do the figures that are widely shared in the media and in policy debates come from? This report delves into the assumptions and uncertainties that are central to these influential calculations. Life cycle assessment models are frequently used, but the data are often derived from a limited set of cases, mostly from industrial systems particularly from Europe and North America. We identify 10 core assumptions and gaps in such assessments. These centre on the limitations and biases of the data; the way systems are analysed – what’s included and excluded; and how baselines are defined and alternatives assessed.