The projects summarised by Room and Rehm and Stockwell et al. are important additions to the evidence base underpinning low-risk drinking guidelines for the general public and to thinking about how such guidelines should be developed, formulated and presented. In this commentary, I will firstly consider the use of guidelines in general terms and defend the proposition that they serve an essential purpose in the attempt to reduce alcohol-related harm. I will then discuss a more specific issue to do with the communication of advice on drinking—how many kinds of recommendation are needed? By coincidence, this special issue of the Drug and Alcohol Review will be published shortly after a consultation by the House of Commons Select Committee for Science and Technology on issues related to drinking guidelines for the UK. At the time of writing, the results of the consultation are unknown, but this commentary provides an opportunity to discuss the relevance of the new Australian and Canadian guidelines to the UK situation.