Optimizing Supply Chain Sales Territories through Incremental Brownfield Design

Abstract: Territory design or districting consists of assigning basic geographic units to territories or clusters according to some relevant criteria. Districting is a tactical decision-making process that has become a fundamental factor within the supply chain. In districting problems, we refer to approaches that construct territory designs from scratch without considering existing plans as Greenfield approaches. However, in business environement, analytics practitioners rarely have the luxury of starting from scratch. Most businesses operate on Brownfield designs-legacy maps that may be decades old but serve as the organization’s playbook. These maps contain nuances, historical relationships, and tribal knowledge that Greenfield approaches cannot fully capture. In other words, the focus is on redistricting decisions, that is, find better plans trying not to deviate too much from the existing plan. This redistricting issue has been studied in some applications such as political redistricting; however, in sales districting, Brownfield approaches are practically non-existent. In this chapter, we present a Brownfield approach to a sales redistricting problem. A solution algorithm based on a series of incremental, high-impact moves is presented. This includes empirical evidence over a set of real-world instances, and a comparison with a Greenfield approach. Through our empirical evaluation of four distinct business regions, we demonstrated that an Incremental Brownfield Algorithm provides significant territory improvements with a minimal number of geographical unit reassignments.