While undertaking research on and performance testing of stormwater products at universities is generally mutually beneficial for both manufacturers and universities and should be encouraged. It is imperative to plan every step of the testing procedure and to fully understand, and not underestimate, the consequences of things going wrong. Even an apparently very simple testing program can experience an multitude of complications and require extensive extra (unbudgeted for) resources. This paper presents a case study on some of the problems encountered and the lessons learned while undertaking performance testing of a proprietary GPT system at the University of the Sunshine Coast. The background of GPTs and how they have been tested in the paste os first outlined. The study methodology is then explained and the selection and verification of the monitoring equipment used in the study is summarized. Details are given of the unforseen problems that occurred before and during testing, and how they were dealt with. Finally, conclusions and recommendations are presented.