This paper reports on an action research project with public sector organisations in the UK which experimented with a range of digital methods (social media data mining, social network and issue network mapping, data visualisation), in order to explore their potential usefulness for the public engagement activities of these organisations. We argue that there is a need for small-scale, qualitative studies of cultures of large-scale, quantitative data like ours, to open up spaces in which to reflect critically on the methods with which such data is produced. However, in the paper we highlight the difficulties we had enacting through action research a commitment to both the potential (which might be seen as the action part of action research) and the problems (which might be seen as the research part of action research) of digital methods. Following Hammersley (2002), we suggest that an equal balance between action and research may always be difficult to sustain, in both action research and the use of digital methods. Despite this, we argue that critical discussion of digital methods needs to extend beyond academic spacesthrough this move, we suggest, we might open up a space in which to reflect on how these methods might be used for the public good.