Is It Time To Review Your Supply Chain Design?

Substantial fuel price increases coupled with other increases, such as an increase in driver pay, can may prompt many executives to consider a change in the number of North American distribution centers they use. Here's an analysis that can help you to evaluate where your firm stands.

In a recent class, while discussing supply chain design and distribution center location, a student asked if the recent run up in diesel prices will have any impact on the optimum number of distribution centers. I replied that the precise answer depends on the specific firm's transportation characteristics, but we may be able to determine the approximate sensitivity using some generic data. Specifically, the question is: Do the current increases in the price of diesel fuel suggest a need to re-evaluate the number and location of a firm's distribution centers?

First, we must determine if diesel fuel costs are really outpacing the other inputs to total logistics cost. By comparing changes in diesel fuel prices since 2000 with the changes in the interest rate, along with warehousing fixed and variable costs, the magnitude of such a differential can be determined. Figure 1 illustrates the year year-over over-year changes in the major cost elements influencing distribution system design. Wage rates have been increasing at a very slow rate (less than 5 percent per year). The warehousing rates, including both facilities and labor, have fluctuated up and down. Interest rates, however, have seen a dramatic drop, followed by a recent spike. Only diesel fuel has seen price increases every year since 2001. Furthermore, diesel has seen three price increases of more than 20 percent in the last five years, while the interest rate has only seen one. Based on this information, it certainly appears that diesel fuel has outpaced the other cost factors influencing distribution network design in recent years.

Since December 2000, diesel fuel prices have risen over 62 percent. Figure 2 illustrates the year year-end fuel prices based on data from Transport Topics. Even more alarming is the recent trend, causing diesel to leap 23 percent since December 2004. Since diesel is the primary fuel source for much of U.S. truck and rail shipping, transportation costs have certainly been impacted. For example, assuming that fuel represents approximately 30 percent of overall transportation cost, a fuel price increase of 25 percent increase overall transportation cost by 7.5 percent. It is reasonable to believe that most distribution networks may not be robust enough to remain optimum over such a wide range of fuel prices.

Second, we must investigate the total logistics costs under various location combination scenarios. As most logisticians know, the logistics total cost model is useful for determining the optimal number of distribution centers for a geographic area such as North America. The trade-offs used in the model include transportation to and from distribution centers, inventory carrying cost, and fixed and variable facility costs. However, due to their relative magnitude, transportation and inventory carrying cost are the major drivers, having the largest impact.

The hypothesis is that distribution network design will be sensitive to 25-40 percent increases in diesel fuel prices. To test the hypothesis, we used LogicNet 6.1(tm) modeling and optimization software produced by Logic Tools, Inc. We used the "Metal Works Case" which is one of the standard examples and replicates the distribution network for a durable product with two plants, 11 product groups, and a combination of truckload and LTL (less than truckload) deliveries. Using the optimizer, we successively identified the total cost curve at a base level, and then increased the fuel cost by 25, 50, and 100 percent respectively. Assuming that 30 percent of transport cost is related to fuel, these fuel cost increases resulted in 7.5, 15, and 30 percent increases in overall transport cost. The simulation results contain representative values for transportation, inventory carrying as well as, facility fixed and variable cost. Table 1 summarizes the total (non-production related) costs for a range of distribution centers for each fuel cost scenario. The minimum cost number of distribution centers is illustrated by the bordered cell. Table 2 illustrates the particular distribution center locations in each network.

The results illustrate the impact of fuel price increases. In the base case, the optimum number of distribution centers is seven. A fuel price increase of 25 percent (i.e., an overall transport cost increase of 7.5 percent) results in no change in the optimum number of distribution centers; however, it is very close to shifting to an optimum of eight distribution centers. A fuel increase of 50 or 100 percent increases the optimum number of distribution centers to ten. In particular, the optimization suggests that distribution centers should be opened in New England (Albany), Northwest (Portland), and a second facility in the Midwest (Indianapolis).

Although these results should not be generalized, particularly across firms that have unique transportation characteristics, they do suggest that optimum distribution system design is sensitive to the fuel price increases that we have seen over the past year. Nevertheless, the 25 percent increase is not enough to force a change. Above the 25 percent level, however, it appears that the optimum number increases by one distribution center for each 25 percent increase in fuel cost. Hopefully, we don't see fuel price increase of that magnitude for a long time.

These results indicate the relative sensitivity of the number of distribution centers to fuel prices. So now I can tell my student that even with the 25 percent increases in fuel over the last year, there should not be any major impact on the number or location of distribution centers. This is undoubtedly due to the fact that 25 per increase in fuel yields a relatively minor (7.5 percent) impact on transport rates. However, a 25 percent increase in fuel in conjunction with a 25 percent increase in driver pay (due to the shortage of drivers) would yield a 15 percent increase in transport rates and would have the same result as a 50 increase in fuel. Under this potential scenario (which is reasonably likely), this analysis suggests that a change in the number of distribution centers would be recommended.

In summary, these results suggest that the combination of fuel price increases and driver wage increases could have some impact on the optimum number and possibly the location of distribution centers to serve North America. However, the impact of the change on total cost may not be particularly significant given the other cost components. Specifically, at the 100 percent fuel increase level, the total (non-production) cost for seven distribution centers is (U.S.) $49,567 USD while the cost at the optimum number of distribution centers is $48,751 USD. The resulting change in non-production costs are is only 1.7 percent, which is probably not substantial enough to justify significant change. So, for my students and the managers who have wondered regarding the nature of the impact, 25 to 50 percent increases in fuel prices and driver wages certainly makes and impact on the total cost but it may not be enough to justify substantial design changes and the required efforts to make the changes.