This article explains the methodological framework created for the ‘Tweeting the Olympics’ project represented in this Journal issue that can be used and adapted by other researchers in their studies of global media events. This project was a case study in the adoption of and adaptation to social media in a global news organisation. It examined the opportunities and challenges that were faced by the BBC World Service during the 2012 London Olympic Games in developing their Twitter strategy, and how audiences responded. The aim of the project was to assess whether and how the BBC World Service engaged audiences via Twitter and fostered intercultural dialogue (or, in the BBC’s terms, a ‘global conversation’). We focused on the uses of Twitter by four different Language Services provided by the BBC World Service: the BBC Arabic Service; BBC.com (the English Language Service targeted at overseas audiences); the BBC Persian Service; and the BBC Russian Service. We adopted a ‘social life of methods’ approach that treats methods as active agents in institutional processes. We argue that social media research and methods, despite their rapid emergence and proliferation, are still at an early stage of development and should be treated as experimental. As we develop methodological designs for our research experiments, sharing our experiences of failure as well as of success is important to advance the field. For although some regard ‘big data’ as the new gold standard, promising forms of knowledge previously unattainable, it is wise to be cautious. There are risks as well benefits for academics working alongside corporate researchers. The paper offers an honest and judicious assessment of the framework we created and used, and suggests new fruitful lines of enquiry.