Since Brumfit's landmark definition of applied linguistics as the theoretical and empirical study of real world problems in which language plays a central role (Brumfit, 2000), there have been periodic calls for applied linguistics to engage with problems experienced by people in real world contexts (such as teaching, health, business, law, social services, business or family), rather than problems of research methodology originating in the research community, and to work to address them, both in policy and practice (Bygate, 2004; Tarone, 2013, 2015; Shuy, 2015; Widdowson, 2017). This principle may well apply to all areas of applied linguistics, but in this piece I would like to explore it in relation to task-based language teaching (TBLT). This is because while TBLT is characteristically defined in terms of the needs and interests of language teachers and learners, it is also informed by research, which is heavily shaped by the priorities of the academy, an influence which can lead it away from some of its real world objectives. Yet if proponents fail to adequately address the priorities and needs of classroom stakeholders, proposals will be doomed to failure, a point acknowledged by many (see inter alia Gatbonton and Segalowitz (1988, 2005), Edwards and Willis (2005), Thornbury and Slade (2006), van den Branden (2006), Eckerth (2008), Andon and Eckerth (2009), Ellis (2009), Gatbonton (2015), Long (2015) and Samuda, Bygate, and van den Branden (2018)). That is, research needs to engage not just with models of second language acquisition (SLA), but with the practices, demands, pressures, and perspectives of stakeholders in real world language classrooms.