Source: adapted from Chicago Consulting.
Freight distribution is a market servicing activity, implying that the nature and structure of the market will have an important impact on how it is serviced. Freight distribution is deeply embedded within its regional geography. Two related variables can help define the level of service performed by a distribution network composed of distribution centers, each covering a specific portion of the territory. The first is the average distance to customers, which defines the market size and the delivery and return (backhaul) distance. The second is the lead time which is the time it takes for a distribution center to deliver the goods once an order has been placed.
The above four maps represent four optimal scenarios in the location of distribution centers within the United States, based upon their number. It is based on two fundamental assumptions. The first is that the market potential is directly related to the accessible population. The second is that deliveries are done by trucking using the highway network. It represents a standard p-median location problem where a p number of facilities must be located to minimize total transportation costs.
For three distribution centers (p=3), the optimal locations would be Allentown, McKenzie, and Bakersfield, handling respectively 31%, 46%, and 23% of the total throughput and an average lead time of 1.29 days. In this case, the throughput is simply a function of the total population within their respective market areas. The other three maps depict the improvements in lead time and average distance to customers if additional distribution centers are considered (5, 7, and 10). After 6 distribution centers, there are limited marginal improvements in lead time (e.g. 1.07 days with 7 distribution centers and 1.04 days with 10 distribution centers).
The frequency of service is an important variable that influences the suitable number of distribution centers servicing a supply chain. A low-frequency level tends to require fewer distribution centers than a high-frequency level. A limited number of distribution centers also implies longer distances for empty returns, which is costly.