Prior studies have demonstrated the importance of bridging structural holes across functional groups in IS projects. In this study, we argue that bridging structural holes is necessary but insufficient for ensuring project success. An additional requirement is that knowledge holes across functional groups need to be bridged to enable effective problem-solving across functional groups. We propose and empirically study the concept of knowledge holes in a case study of an ERP upgrade. Our findings suggest that complementary to the concept of structural holes, the concept of knowledge holes is useful for explaining different project outcomes. Our findings also demonstrate methods for bridging knowledge holes. Contributions of this study are manifold.