Estimating animal abundance is an area of interest for many conservationists and population ecologists. When using count data to obtain abundance estimates, issues of imperfect detection must be taken into account. The N-mixture model proposed by Royle (Biometrics, 60(1), 108–115, 2004) provides a solution to this issue by incorporating detection probability in the estimate of abundance. We examine the original N-mixture model and assess its uses, advantages and assumptions. We then examine extensions to this vanilla N-mixture model which allow for the estimation of abundance using a range of data unsupported by the original model, including data that contain observations of multiple animals, data that contain large numbers of zero-counts and data that are collected over long periods of time. We finish by illustrating the applicability of both the original N-mixture model and a model proposed for multiple species collected over several years (Mimnagh et al., Environmental and Ecological Statistics, 1–24, 2022) by estimating foraging populations for bee species from data collected in 2016 and 2019 as part of the BeeWalk Survey in the UK.