Camera traps are increasingly used to survey and monitor rare or cryptic species, yet few studies consider how camera orientation influences species detectability, among other metrics such as total independent detections and likelihood of missing detections. We used these measures to compare the performance of vertically and horizontally orientated camera traps at 46 sites spread over 10 000 km2 in north-west Australia. Data were collected for four taxa, including northern quolls (Dasyurus hallucatus), Rothschild's rock-wallabys (Petrogale rothschildi), feral cats (Felis catus) and varanids (Varanus spp.). Metrics compared included probability of species presence/absence, total independent detections recorded and likelihood of cameras missing or recording nightly detections. We found camera orientation did not impact camera performance across any metric for northern quolls. By contrast, we found horizontal cameras were more efficient at detecting feral cats and Rothschild's rock wallabies. They also recorded more detections and were less likely to miss detections than vertical cameras for these species. For varanids, vertical cameras outperformed horizontal cameras across all metrics. Studies that use vertical cameras to collect image data better suited for species or individual identification should consider how target species detectability may be compromised by having a reduced detection zone size. However, horizontally orientated cameras may not always be superior to vertically orientated cameras in terms of species detectability, particularly for laterally compressed species such as lizards.