Reducing all water loss components to zero is neither technically possible nor economically viable. The water loss components should be accurately assessed and prioritised for a reduction. This paper investigates all methods that break down the water losses in distribution networks into apparent and real losses. Their accuracies and uncertainties are discussed and applications to three case studies in developing countries are presented. The results show that different methods estimate the water loss components differently. Consequently, different reduction measures are planned and prioritised. Interestingly, the least accurate methods have a low level of uncertainty, but more realistic assumptions yield higher uncertainties. This suggests that the uncertainty analysis only assists in improving the outputs of each of the methods but does not demonstrate their accuracy. The cost of water loss varies depending on the used assessment method and the economic feasibility of the reduction measures is significantly influenced. The water loss components should therefore be assessed for the whole network using at least two methods to reasonably model and monitor the loss reduction in water distribution networks.