The combination of conflict, food insecurity, and displacement generates competing claims for financial resources that stretch the donors' ability to provide funding and the humanitarian organizations' capacity to provide social assistance. The paper uses Receiver Operating Characteristic curves and related indexes to determine the optimal targeting strategy of a food voucher program for refugees. The estimations focus on the 2014 food vouchers administered by the World Food Programme to Syrian refugees in Jordan. The analysis uses data collected by the United Nations High Commissioner for Refugees. Based on a poverty model, Receiver Operating Characteristic curves are used to optimize coverage and leakage rates under budget constraints. The paper shows how policy makers can use these instruments to fine-tune targeting using coverage rates, budgets, or poverty lines as guiding principles to increase the overall efficiency of a program. As humanitarian organizations operate under increasing budget constraints and increasing demands for efficiency, the proposed approach addresses both concerns.