Abstract Generalisation and specialisation in species‐species interactions are key ecological concepts for interpreting the different interaction patterns observed in nature. Hence, finding the best way to operationalise them has been a major quest in Ecology. This quest has led to considerable conceptual development, and now the observed interaction pattern of a species is assumed to be a combination of three factors: its degree of generalisation, abundance‐driven neutral effects, and sampling effects. Here, we aimed to assess the influence of these factors on the performance of previously proposed indices of generalisation. To do so, we used simulated data that allowed us to separate and analyse independently the influence of each factor. Our assessment shows that the estimates made by most traditional indices are affected by differences in resource abundance distribution, leading to over‐ or underestimation of how generalised a consumer is. To solve this problem, we propose a new index that remains unaffected by neutral effects and is robust to sampling effects. Our new index may help to understand what interacting species require to keep viable populations and how they might respond to changes in resource availability.