A challenge in Information Systems Design Science Research (ISDSR) is the evaluation of constructed Information Systems Design Theories. March and Smith (1995, p. 254) state that “significant difficulties … result from the fact that artifact performance is related to the environment in which it operates.” Therefore, it is nearly impossible to achieve stable knowledge of the utility of ISDSR artifacts. This is a severe problem because—as in each scientific discipline—it should be able to show that there is progress in ISDSR. In this paper, we propose an approach for improving the stability of knowledge of the utility of ISDSR artifacts. We focus on instability resulting from the social setting of an evaluation context. By applying Habermas’s Discourse Theory, we show why a separation of stakeholder types is essential for gaining stable knowledge on the utility of ISDTs.