Despite the consistent claim that companion animals can and do alleviate human loneliness, a recent systematic review of quantitative studies of human loneliness and companion animals (Gilbey and Tani 2015) found no evidence to support this ‘belief’ (as they put it), except in animal-assisted therapy (and even there the authors were not entirely convinced that they do). Taking their article as a starting point this paper develops a critical examination of quantitative methodologies that have been used to date and suggests that they have not taken into account the extent and complexity of contemporary human loneliness or how companion animals configure with it. This paper argues that recent advances in the understanding of human-animal relations suggests that a new form of qualitative study is needed that can focus on how relationships between companion species and humans are created and in what way these relationships impact on human loneliness. A new auto-ethnographic case study is provided in order to indicate the method and its potential value.