Urban landscapes present substantial opportunities for biodiversity conservation with residential gardens offering some of the greatest potential conservation gains given that they represent a significant proportion of the total greenspace in urbanised landscapes. However, knowledge of wildlife ecology within gardens remains scarce, likely due to the difficulties associated with field ecologists accessing privately owned areas. Citizen-sourced data presents an alternative approach to typical field-based investigations and could be used to develop an extensive understanding of biodiversity within privately owned green spaces, providing concerns of data unreliability could be overcome. Here we i) examine the potential for urban householders to identify a threatened nocturnal mammal, in the presence of a similar non-threatened species, ii) examine which attributes can predict the level of accuracy in householders’ species identifications and iii) investigate how their self-reported level of certainty affects identification reliability. We found up to 80% agreement between householders and experienced ecologists when we assessed presence and absence of nocturnal mammals on urban properties. The amount of time a householder had lived at their residence was a significant predictor of accuracy for mammal species identification, suggesting that familiarity with a site enhances the accuracy of citizen science data. Those people with a high level of certainty in their ability to correctly identify nocturnal mammals were no more likely to show higher species identification accuracy than those with low certainty. In urban areas, where ecological surveys for nocturnal taxa are especially challenging, our results inspire optimism that householders can add to the body of knowledge about biodiversity persisting in these landscapes, especially on properties where they have resided for at least one year.