Due to the increased importance and usage of self-managed virtual teams, many recent studies have examined factors that affect their success. One such factor that merits examination is the configuration or composition of virtual teams. This article tackles this point by (1) empirically testing trait-configuration effects on virtual team performance, which are based on supplementary, complementary and interaction person-team fit perspectives and (2) extending the suggested trait-configuration model to include virtual team configuration in terms of the perceived problem-solving demands of the task as a predictor of team performance. To this end, median regression techniques were applied to data from 62 self-managed virtual teams that used an asynchronous bulletin board for working on a case study analysis. The findings suggest a plausible negative main effect of within-team conscientiousness heterogeneity on team performance, operationalised as standardised team grade. This effect depends on the level of the within-team extroversion heterogeneity which helps to mitigate the negative effect of within-team conscientiousness heterogeneity on team performance. Furthermore, within-team heterogeneity of the perceived problem-solving demands of the task reduces team performance. Overall, this study proves that virtual team configuration matters, and demonstrates that the joint utilisation of multiple person-team fit perspectives for improving virtual team performance has merit. Implications for research and practice are further discussed. (Contains 2 tables and 1 figure.)