Many potential new medicines fail in phase III clinical trials, because of either insufficient efficacy or intolerability. Such failures may be caused by the absence of an effect and also if a suboptimal dose is being tested. It is thus important to consider how to optimise the choice of dose or doses that continue into the confirmatory phase. For many indications, it is common to test one single active dose in phase III. However, phase IIB dose-finding trials are relatively small and often lack the ability of precisely estimating the dose–response curves for efficacy and tolerability. Because of this uncertainty in dose response, it is reasonable to consider bringing more than one dose into phase III. Using simple but illustrative models, we find the optimal doses and compare the probability of success, for fixed total sample sizes, when one or two active doses are included in phase III. Copyright © 2012 John Wiley & Sons, Ltd.