<p>This paper reports on a study examining the use of ABRACADABRA (ABRA), a Canadian web-based tool for supporting early literacy instruction that was trialled in the Northern Territory of Australia over the period 2008–2010. The three year trial established ABRA's effectiveness in urban and remote primary schools with a mix of Indigenous and non-Indigenous students under quasi-experimental and experimental conditions. Both this Australian trial and preceding studies in Canada demonstrated ABRA's capacity to generate significant student outcomes against a range of literacy measures. These studies further found student effects are greatly enhanced when teachers confidently integrate ABRA content into their broader literacy program" and conversely, that ABRA has reduced impact when teachers are less confident with integrating the technology into their teaching. Given ABRA is freely available on the internet, we additionally felt it was important to consider ABRA's likely implementation fate in non-research circumstances. The study reported here examines four north Australian primary schools which implemented ABRA outside of trial conditions, and was conceived as something of a pre-emptive strike against premature uptake of this otherwise promising program. We develop our analysis from classroom observations and interviews with practitioners, and explore how ABRA might fare if it were implemented with minimal support" or rather, with a level of support equivalent to that typically offered in Northern Territory schools for other literacy programs. Our findings confirm a universal education truism about the importance of carefully targeted training and support to ensure optimal outcomes for program effect" a truism which arguably has greater import in the turbulent school environments facing socially disadvantaged students in north Australian schools. This study has implications for how educational interventions, particularly in remote and cross-cultural settings, might be implemented and sustained at scale.</p>