The first phase of development of the Australian Water Resources Information System (AWRIS) focussed on developing the means to receive, harmonise and then publish water storage observations as a consistent national dataset. This paper describes the approach that was taken in developing the first AWRIS operational system. It discusses what we have learned about the challenges of harmonising data from a diverse array of sources; identifies the key lessons we have learned, and highlights how we will be modifying our approach in the next AWRIS development phase.