Spiders are dominant predators in terrestrial ecosystems, regulating invertebrate biomass through predation. However, analysing predation patterns in the field is challenging, especially as spiders are extra-intestinal liquid feeders and often active at night. Gut content metabarcoding (GCM) provides a cost-effective method to study the prey composition of spiders. In this study, we determine the minimum sample size of spider specimens required to obtain representative estimates of local prey composition. Six common spider species were studied using GCM and we show that the number of analysed individuals per spider species influenced the saturation levels for prey at different taxonomic levels. To achieve 90% saturation at the prey species level, between 50 and over 150 individuals per spider species had to be analysed, which is challenging for ecological multi-site studies. Hence we argue for consideration of coarser taxonomic levels, such as family or order, which often provide sufficient ecological information to understand predator-prey interactions. At these levels, saturation can be achieved with much smaller sample sizes of 4 to 8 individuals per species, making ecological GCM studies more cost-effective. These results highlight the benefits of considering the taxonomic resolution in metabarcoding studies to address ecological research questions. We further provide information that will facilitate future investigations of predator-prey dynamics, not just in spiders, but also in other extra-intestinal liquid feeders like for example ground beetles.