Social network analysis (SNA) has become a mature scientific field over the last 50 years and is now an area with massive commercial appeal and renewed research interest. In this paper, we argue that new methods for collecting social nework strucuture, and the shift in scale of these networks, introduces a greater degree of imprecision that requires rethinking on how SNA techniques can be applied. We discuss a new area in data management, probabilistic databases, whose main research goal is to provide tools to manage and manipulate imprecise or uncertain data. We outline the application building blocks necessary to build a large scale social networking application and the extent to which current research in probabilisitc databases addresses these challenges.