We propose a strategy for implementing genomic selection in plant breeding programs to develop inbred lines that reorganizes traditional breeding programs into two distinct components. These components are: (i) a population improvement component to develop improved germplasm through rapid recurrent selection; and (ii) a product development component to identify new inbred varieties or parents for hybrids using traditional breeding program designs. Stochastic simulations of entire breeding programs over 40 years were used to evaluate the effectiveness of this strategy relative to a conventional program without genomic selection and programs using three standard strategies of implementing genomic selection. Cost-effectiveness was measured by constraining all programs to approximately equal annual operating costs and directly comparing each program’s overall performance. Programs using the two-part strategy generated between 2.36 and 2.47 times more genetic gain than the conventional program and between 1.31 and 1.46 times more genetic gain than the best performing standard genomic selection strategy. These results indicate that the two-part strategy is a cost-effective strategy for implementing genomic selection in plant breeding programs.