Abstract Water-using processes are typically modeled as either fixed flowrate operations or fixed contaminant load operations. A new method for targeting the minimum freshwater and pinch in a single-contaminant water network is proposed, which can be applied to both kinds of operations. The method consists of plotting separate source and demand composites with flowrate as the horizontal axis and contaminant load unusually as the vertical axis. It is elegant, non-iterative, and can handle hybrid problems where both kinds of operations coexist. To design minimum freshwater networks for fixed flowrate problems, an algorithm is presented based on a newly developed principle of nearest neighbors. The principle simply states that the source streams to be chosen to satisfy a particular water demand must be the nearest available neighbors in terms of contaminant concentration. To design minimum freshwater networks for fixed contaminant load problems, the nearest neighbors algorithm is applied to process units that lie across the pinch. Units that lie entirely on one side of the pinch are satisfied by the cleanest source available on that side of the pinch. In other words, below-pinch units are satisfied by freshwater and above-pinch units are satisfied by the cleanest available stream above the pinch. Designs based on this methodology, apart from meeting the minimum freshwater target, also minimize the water flowing through the process units resulting in reduced network capital cost.