Social navigation is a promising approach for supporting privacy and security management. By aggregating and presenting the choices made by others, social navigation systems can provide users with easily understandable guidance on security and privacy decisions, rather than requiring that they understand low-level technical details in order to make informed decisions. We have developed two prototype systems to explore how social navigation can help users manage their privacy and security. The Acumen system employs social navigation to address a common privacy activity, managing Internet cookies, and the Bonfire system uses social navigation to help users manage their personal firewall. Our experiences with Acumen and Bonfire suggest that, despite the promise of social navigation, there are significant challenges in applying these techniques to the domains of end-user privacy and security management. Due to features of these domains, individuals may misuse community data when making decisions, leading to incorrect individual decisions, inaccurate community data, and "herding" behavior that is an example of what economists term an informational cascade. By understanding this phenomenon in these terms, we develop and present two general approaches for mitigating herding in social navigation systems that support end-user security and privacy management, mitigation via algorithms and mitigation via user interaction. Mitigation via user interaction is a novel and promising approach to mitigating cascades in social navigation systems.