A simple and direct method for processing residence time distribution (RTD) signals from conductivity data was developed for a spiral wound membrane reverse osmosis (RO) system. Two models were tested: axial dispersion (AD) model and exponentially modified Gaussian (EMG) model. The results show that the present method provides a simple, fast and accurate RTD data reduction. The presence of fouling in the small laboratory RO module increased the axial dispersion coefficient, indicating that fouling in the membrane channel augments the axial dispersion extent, thus resulting in the deviation of flow pattern from ideal plug flow. The dead volumes gradually decreased with increasing Reynolds number for the worn industrial module, while there was no dead volume for the new industrial module. As a result, the membrane wearing-out could increase the dead zones in the industrial spiral wound module.