In order to reduce the number of operations for the assessment of potable water treatment, principal component analysis and hierarchical clustering are applied to large databases of raw and treated water of three treatment plants with various processes. It appears that the measurements can be divided into three clear groups, with a correlation higher than 0.8. The first contains salinity, conductivity, water hardness, calcium, magnesium and chlorides. The second includes turbidity and organic matter. The third includes pH and alkalinity. Despite the disparities in water quality and in all the cases, three parameters were sufficient to represent all the routine measurements: conductivity, turbidity and pH, which can represent the three principal components of the data. It can reduce by two-thirds of the measurement and analysis, dropping from 6,960 to 2,088 analysis annually. The analysis on the principal axes of the individuals, represented by raw and treated water from the three treatment plants, reveals that the quality of the raw water seems more important than the type of treatment process, in the resulting quality of treated water. These results could be generalized and easily adopted by other treatment plants whatever the process. They could offer substantial savings of time, chemicals, electricity and longevity of the devices.