Product change in glass furnaces is a common and challenging industrial problem generally associated with long transition times and possible ream defects. A reduction in the product turnover time implies a higher product yield and a reduced energy consumption. This turnover time relies to a large extent on the understanding of furnace dynamics. In this work, the disadvantages of the current transition practice (based on the concepts of minimum residence time and perfect mixing) are discussed. An optimization approach that is based on the residence time distribution of the glass furnace is tailored to accommodate practical implementation issues in the product transition control. Comparisons of the current transition practice and the proposed approach are made through computer simulations, which demonstrate that significant savings in the product turnover time are expected if the latter is used.