The supervisor of a graduate research candidate has a responsibility to the candidate to provide guidance throughout the project. It is necessary to ensure that the candidate accumulates a comprehensive understanding of the research endeavour, delivers a dissertation that evidences that understanding, and makes a sufficient contribution. Through that process, the candidate achieves 'rite of passage' to a research career. On the other hand, the supervisor also has a responsibility to themselves and their profession: to sustain their own integrity. The Information Systems discipline has been dragged towards a state in which rigour dominates relevance to an untenable extent, and the real-world value of too much of the research that is conducted and published is diminishingly small. Nowhere is the tension so great as in the research domain that is the focus of the Bled eConference. The dilemma that I expand on in this paper is created by the conflict between the current expectations of the discipline and the needs of real-world-relevant research.