In a growing number of countries, governments and public agencies seek to systematically assess the scientific outputs of their universities and research institutions. Bibliometrics indicators and peer review are regularly used for this purpose, and their advantages and biases are discussed in a wide range of literature. This article examines how three different national organisations produce journal ratings as an alternative assessment tool, which is particularly targeted for social sciences and humanities. After setting out the organisational context in which these journal ratings emerged, the analysis highlights the main steps of their production, the criticism they received after publication, especially from journals, and the changes made during the ensuing revision process. The particular tensions of a tool designed as both a political instrument and a scientific apparatus are also discussed.