Developing and validating methods to determine trends in populations of threatened species is essential for evaluating the effectiveness of conservation interventions. For cryptic species inhabiting remote environments, this can be particularly challenging. Rock wrens, Xenicus gilviventris, are small passerines endemic to the alpine zone of southern New Zealand. They are highly vulnerable to predation by introduced mammalian predators. Establishing a robust, cost-effective monitoring tool to evaluate population trends in rock wrens is a priority for conservation of both the species and, more broadly, as part of a suite of indicators for evaluating effectiveness of management in New Zealand's alpine ecosystems. We assessed the relative accuracy and precision of three population estimation techniques (mark-resight, distance sampling and simple counts on line transects) for two populations of rock wrens in the Southern Alps over six breeding seasons (2012-2018). The performance of these population estimators was compared to known rock wren population size derived from simultaneous territory mapping. Indices of abundance derived from counts on transects were correlated with territory mapping at both study areas, and performed better than either mark-resight methods or distance sampling. Simple counts on standardised line transects are a highly cost-effective method of monitoring birds because they do not require banding a population. As such, we recommend that line transect counts using the design outlined in this paper be adopted as a standard method for long-term monitoring of rock wren populations. Although species-specific testing is required to validate use of low-cost population indices, our results may have utility for the monitoring of other cryptic passerines in relatively open habitats.