This paper examines how processes of ‘datafication’ associated with contemporary education accountability mechanisms, including high-stakes National Assessment Program – Literacy and Numeracy (NAPLAN) testing in Australia and valued-added measures (VAM) of teachers’ effects on student achievement data in the USA, are reconstituting not only teaching practice but also, and perhaps more importantly, the teacher subjects themselves. In recent years, there has been a significant increase in data collection and use in education globally, with a specific reliance on numbers for measuring and governing performance (Ozga, 2009). Consequently, there is a growing sense of policy and governance by numbers (Grek, 2009; Lingard, 2011; Rose, 1999) that is enabling teacher performance to be steered in particular ways. While putatively ‘objective’ and numerical measurements are used to legitimate school or system-level policy decisions, they can also influence how individual teachers come to enact and value certain pedagogical practices over others (Hardy & Lewis, 2016; Lewis & Hardy, 2015). Drawing suggestively across literature and theorising around educational performativity (Ball, 2003), governing by numbers (Rose, 1999) and the constitutive power of numbers to ‘make up’ people (Desrosières, 1998; Hacking, 1985), we examine how enumerations of teaching performance at two schools in Australia (via NAPLAN) and the USA (via VAM) function to discursively constitute the teacher as a performative subject. Our analyses are informed by interviews with teachers and school leaders, as well as relevant school-level policy documents. We argue that such accountability practices render teacher ‘quality’ as something visible, measurable and comparable, remaking the ‘teacher’ into a calculable object of knowledge and, significantly, limiting the conditions of possibility (Foucault, 1994) by which teaching ‘excellence’ might otherwise be conceived. Moreover, the representation of data-informed teaching performance, such as improving ‘distance travelled’, were at times prioritised over other more educative and substantive practices, including (at times) more authentic student learning, even while these very practices were often problematised by teacher themselves. As such, we see such developments as reflecting the transformation of the teaching profession into, arguably, a ‘profession’ of data, in which ‘re-presenting’ (Miller & Rose, 2008) oneself as an effective teacher through data has become of central importance, both for system-level accountabilities and, indeed, for expressing one’s own worth as a teacher. However, and at the same time, the presence of teacher contestation with these acts of data-informed ‘profession’ suggests a way towards more productive responses to such practices.