Many government responses to the coronavirus-pandemic have been marked by attempts at expertization and scientization. Particularly, politico-epistemological authority is being given to the behavioural science community consulting government. This article critically scrutinizes this most recent wave of behavioural expertization. Taking developments in the UK and the Netherlands as our case-studies, we shed light on the disparate ways in which behavioural expertise is being (re)shaped during COVID-19. Some of these ways point at processes of behavioural expertise ‘drift’, in which the applicability and robustness of this knowledge source gets overstated. Other ways instead point at processes of behavioural expertise ‘thrift’ or ‘shift’, where the knowledge is used only minimally or taken in wholly new and norm-breaking directions. Doing so, we seek to demonstrate the importance of institutional context in understanding how behavioural expertise is currently shaping public policy: underpinning institutional configurations determine whether the expertise is gauged and applied effectively.