Real-time monitoring of distribution water networks relies on the deployment of sensors and the availability of their measurements in order to predict the system state and assess its performance. A meaningful application of this methodology is the detection and localization of leaks using model-based approaches. Since the number of sensors is limited because of budget constraints, it is important to place these devices in locations where the effectiveness of the leakage diagnosis is maximized. Finding the best sensor distribution is a global optimization problem defined by an objective function that might depend on different factors. Therefore, deriving the correct structure of such function is a crucial step as a wrong definition would lead towards a confusing optimal solution affecting negatively the monitoring performance. In general, sensor placement optimization methods describe objective functions using factors related to the amount of undistinguishable leaks. More concretely, the methods first compute groups of locations where leaks cannot be differentiated and then maximize this number of groups or minimize their size. In this paper, additional factors are presented to accurately represent the requirements of the leak diagnosis phase. These include other statistical figures related to the size of groups, geographical characteristics like the group’s extension area, levels of sensitivity that indicate whether a location is more or less sensible to pressure changes, etc. The objective of this study is to review several factors in order to comprehend their behaviour and justify or discard them for the objective function. The indicators under study are evaluated by means of a cross-correlation analysis applied to the scenario defined by the District Metered Area of the Barcelona water distribution. Results indicate the existence of different independency levels between the indicators that allow us to select those with less redundancy.