The methodology of investment appraisal through cost-benefit analysis has received a lot of attention in the economics literature. However, the major part of the work has focused on the evaluation of yearly benefits and their components. Most guidelines list some indicators, such as the internal rate of return, the net present value per euro spent or the first-year benefit-cost ratio, but with no clear indication of how and in which cases they should be used. Building on the transport case, this paper aims to fill this gap. It recalls the teachings of economic analysis for project prioritisation, based on the principle of maximising the net present value. Through simulations, it examines how the various indicators perform, considering two situations, with and without budget constraints. It turns out that the various indicators perform roughly the same as long as the projects under comparison have similar benefit patterns over time, and using them does not induce losses that are too large, though the order of implementation is rather different from the optimal one. However, the indicators lead to huge under-optimisations when the projects have different patterns. The conclusion is that rigorous programme optimisation should be preferred to the use of indicators, and would encourage a wider use of cost-benefit analysis, which is presently too often limited to checking projects one by one.