ABSTRACTProduct development is typically challenging. There is a strong desire to understand critical performance factors early in order to optimize formulations for Phase III trials or marketing approval. However, over 90% of drugs entering development in humans do not make it to the market. Thus there is a tremendous risk that resources spent early in development will be wasted as candidates attrite. To overcome this apparent dilemma, a Bayesian approach is suggested. In ‘traditional’ product development, clinical study designs are typically relatively assumption free and often do not consider learning from earlier investigations. A Bayesian approach is one in which prior information is considered when interpreting the results of the current study. A candidate IVIVC can be developed at the earliest stage of clinical development. Subsequently there are multiple opportunities within a typical development program to assess the performance of the preliminary IVIVC as new formulations are introduced. With each introduction, there is an opportunity to either confirm the adequacy of the IVIVC or to identify the need for development of a more predictive dissolution test and IVIVC. Copyright © 2012 John Wiley & Sons, Ltd.