The gut microbiomes of birds and other animals are increasingly being studied in ecological and evolutionary contexts. Numerous studies on birds and reptiles have made inferences about gut microbiota using cloacal sampling; however, it is not known whether the bacterial community of the cloaca provides an accurate representation of the gut microbiome. We examined the accuracy with which cloacal swabs and faecal samples measure the microbiota in three different parts of the gastrointestinal tract (ileum, caecum, and colon) using a case study on juvenile ostriches, Struthio camelus, and high-throughput 16S rRNA sequencing. We found that faeces were significantly better than cloacal swabs in representing the bacterial community of the colon. Cloacal samples had a higher abundance of Gammaproteobacteria and fewer Clostridia relative to the gut and faecal samples. However, both faecal and cloacal samples were poor representatives of the microbial communities in the caecum and ileum. Furthermore, the accuracy of each sampling method in measuring the abundance of different bacterial taxa was highly variable: Bacteroidetes was the most highly correlated phylum between all three gut sections and both methods, whereas Actinobacteria, for example, was only strongly correlated between faecal and colon samples. Based on our results, we recommend sampling faeces, whenever possible, as this sample type provides the most accurate assessment of the colon microbiome. The fact that neither sampling technique accurately portrayed the bacterial community of the ileum nor the caecum illustrates the difficulty in noninvasively monitoring gut bacteria located further up in the gastrointestinal tract. These results have important implications for the interpretation of avian gut microbiome studies.