Establishing in vitro dissolution equivalence has become a critical component of process development, post-approval comparability, and biowaver justification. Following a brief critique of the underlying rationale and current practice of dissolution profile similarity testing, we set the stage for a Bayesian approach. We emphasize the importance of a rigorous definition of similarity in the context of an appropriate inference space. We show how Bayesian hierarchical modeling offers an accessible and coherent framework for risk-based similarity decision making. We review a number of possible approaches including the F2 parameter (a link to the commonly used f2 statistic), multivariate and composite hypotheses involving absolute mean profile differences, continuous time profile model parameter differences, and predicted difference in time to Q (the minimum compendial % dissolution requirement). The operating characteristics of the proposed methodologies are determined by computer simulations and several real-data examples are provided to illustrate these approaches.