As family group decision-making (FGDM) is a relatively new child welfare intervention, research has not adequately explored its implementation in multiple contexts or the degree to which implementation matches the previously defined goals, principles, and structures of the model. This article outlines the key components of FGDM and discusses the findings from a two county study of FGDM. Data collected from multiple sources participating in California's Waiver Demonstration Project provide a multidimensional look at model fidelity. Using bivariate analysis and comparing results from multiple sources suggested that although basic FGDM elements were followed, certain philosophies, including community representation and mobilization of supports, were not consistently implemented. In addition, although everyone shared favorable impressions of the conferences themselves, conference participants were more positive than child welfare workers and caregivers about the conference's effectiveness for helping the child. Even if FGDM is implemented as intended, these findings suggest this intervention may not adequately engage family and community members to sustain long-term action which would impact the intended child welfare outcomes. Given the rise in the number of communities utilizing FGDM, measuring intervention fidelity and ultimately linking fidelity to outcomes are important for implementation on a larger scale.