Solar PV-based autonomous mini-grids represent an economically affordable and robust electrification option for rural communities. However, the initial investment cost for renewable energy technologies such as solar PV remains high for rural communities. Implementation of demand-side management (DSM) could increase the cost-efficiency of mini-grids in rural areas. This requires demand-side knowledge, but little is still known of electricity demands in recently electrified areas and, in particular, of how DSM implementation could impact mini-grids. The few studies available focus either on systems or on appliance levels while this study aims to determine cost-efficiency impacts of DSM implementation at a category level. A shifting strategy is applied based on classification into high priority loads and low priority loads. Autonomous rural mini-grid components sizing for four different load categories and load flexibility are carried out using particle swarm optimization. The results show that different load category combinations result in large variations in terms of possible levelized energy cost reductions and, thus, in terms of the cost-optimal sizing of the mini-grid components. The DSM implementation on the household and productive use categories have the largest capacity of reducing the levelized energy cost, by 45.8% and 20.7%, respectively, compared to the no demand-side management case.