The distribution of the merchandise for the retail companies concentrates approximately 30 percent of the operation costs. Costs generally increase if the company wants to have a high level of customer service, that is, achieve the fastest delivery. We focus on improving the distribution system of an automotive spare parts company to deliver its products, offering a high level of customer service. The company has a finite number of delivery drivers, which are assigned to specific areas of the city. Therefore, as a first step, a classification, grouping and assignment of the clients to the drivers is carried out using clustering algorithms, such that a balance of the workload is obtained. Once the coverage areas have been established, it is necessary to calculate departures and schedules in order to organize the visits of the delivery drivers to the customers. Finally, the best delivery routes are calculated through the use of metaheuristics considering the necessary restrictions that improve service times and reduce distribution costs. The results of the methodology applied in the company and a comparison of the savings generated against the current process are shown and cycles of continuous improvement are established.