A primary goal in the design of wireless sensor networks is lifetime maximization, constrained by the energy capacity of batteries. One well known method to reduce energy consumption in such networks is message aggregation. This method reduces the number of messages transmitted in the network, thus extending its lifetime. In order to be effective, aggregation requires messages to be delayed in their path throughout the network. This technique, therefore, must define where and for how long a message should be delayed. This paper shows how aggregation points and corresponding aggregation delays can be determined so that the overall energy efficiency of the system is improved. An algorithm is presented which allows the efficient computation of aggregation points and delays. The benefits of the algorithm is verified through simulation experiments.