In Maharashtra state, India, 25% of the total electricity generated is used in irrigation. The electricity supply for irrigation is of poor quality resulting in low energy efficiency. Most farmers are marginal and small. Only 25% of farmers have access to irrigation. Increasingly unreliable rainfall is leading to a greater requirement for irrigation access. In this work, we investigate two community demand-side management initiatives that can reduce infrastructure requirements, improve energy efficiency as well as the quality of supply to farmers. One is load management, where 25 to 30 farmers connected to one Distribution Transformer, coordinate their pumping schedule, so as not to overload the network. The other is each farmer installing a capacitor on their pumps. The challenge is that they are effective only if all or at least a majority of the group commit to action. A pilot project was conducted successfully among three farmer groups in different villages during the peak irrigation season. Schedules were created considering biophysical, operational, and system constraints. For example, wheat is more sensitive to timely irrigation and was given priority over cotton. While trust is key in this exercise, the limiting factor for these groups is automated schedule generation so that they may conduct it on their own. While the load management initiative required farmers to trust each other and adjust to a schedule, capacitor usage has substantial benefits, for a small expenditure and no downsides. However, since capacitors are not used on pumps, the benefits were unknown to them and it was hard to convince them to install capacitors. Whereas the potential benefits of load management were clear to them from the difference in grid performance between peak and off-peak seasons. Hence besides awareness building, a structured demonstration of capacitor usage may be necessary.