Real-time monitoring of water consumption activities can be an effective mechanism to achieve efficient water network management. This approach, largely enabled by the advent of smart metering technologies, is gradually being practiced in domestic and industrial contexts. In particular, identifying water consumption habits from flow-signatures, i.e., the specific end-usage patterns, is being investigated as a means for conservation in both the residential and nonresidential context. However, the quality of meter data is bivariate (dependent on number of meters and data temporal resolution) and as a result, planning a smart metering scheme is relatively difficult with no generic design approach available. In this study, a comprehensive medium-resolution to high-resolution smart metering program was implemented at two nonresidential trial sites to evaluate the effect of spatial and temporal data aggregation. It was found that medium-resolution water meter data were capable of exposing regular, continuous, peak use, and diurnal patterns which reflect group wide end-usage characteristics. The high-resolution meter data permitted flow-signature at a personal end-use level. Through this unique opportunity to observe water usage characteristics via flow-signature patterns, newly defined hydraulic-based design coefficients determined from Poisson rectangular pulse were developed to intuitively aid in the process of pattern discovery with implications for automated activity recognition applications. A smart meter classification and siting index was introduced which categorizes meter resolution in terms of their suitable application.