Reintroduction and translocation programs are widely used conservation tools but their success rates are low. Poor success rates for reintroduction programs are commonly attributed to insufficient knowledge of species habitat requirements, especially if they are critically endangered. Yet conservation managers are frequently required to make decisions about suitable reintroduction sites when information is incomplete or uncertain. A widely used strategy to assist the selection of reintroduction sites - habitat suitability models - may rely on assumptions and simplifications to fill gaps in existing data. It is essential that these models are then evaluated and refined as new evidence becomes available. In this study, we examine the effectiveness of a reintroduction program based on habitat suitability modelling: that for the critically endangered eastern barred bandicoot Perameles gunnii in Australia. After collecting a variety of data from the reintroduction sites, we found that habitat preferences for this species could be accurately predicted using a simple logistic regression model within two predictor variables rather than the five previously used to select reintroduction sites. This made it possible to better focus limited resources on the most suitable reintroduction sites. We believe that building such a process of review into a reintroduction program can contribute to improving its success, while ensuring that scarce conservation resources are used more effectively.