In a complex, interconnected world, being able to confirm the identity of a person or object is critical. Applications from border control to banking rely on the notion that a system can weed out deception while permitting legitimate access. This trade-off is at the heart of identity discovery. As more scanning technologies become available, data potentially useful to identity discovery is growing exponentially. Different features and processing algorithms can have complementary strengths and weaknesses, so combining them increases identity discovery accuracy. We believe fusion-the use of multiple technologies, algorithms, or modalities in identifying or authenticating people and items-is the better approach because it can improve a system's accuracy, efficiency, and robustness. A simple linear combination fusion method can work almost as well as more analytically sophisticated fusion methods. A framework for generalizing fusion-based identity discovery helps designers think about an identity discovery system in system engineering terms.