Supply Chain Frontiers issue #9. Read all articles in this issue.
Transportation has become a more expensive item on the balance sheet, encouraging companies to be more judicious when using these services. A store delivery model developed by Master of Engineering in Logistics (MLOG) graduate, Chris Kerslake, is helping a US supermarket chain to streamline its delivery services.
"We wanted to put more science behind the selection of delivery days," explained Kerslake. His MLOG thesis project analyzed the frequency of deliveries from a central distribution center to three supermarkets in the chain. He looked at shipments of dry grocery product covering about 10,000 SKUs to three stores that differ in size, sales volumes and delivery schedules. There were three key assumptions: that one store's delivery schedule has no effect on the other stores, holding costs at the DC and stores were identical, and the costs of storing excess product in the backroom and making a single delivery can be quantified.
The replenishment process at the supermarkets is driven by forecasting software that uses sales history data to determine projected sales. Delivery schedules are based on average weekly sales, and the software recommends order quantities using the historical information. The system has a number of drawbacks. Supermarket managers tend to favor more deliveries, while transportation center managers are usually more interested in minimizing the number of times they have to dispatch trucks to stores. A compromise is reached and the delivery days fixed. This traditional procedure is relatively rigid in that once set, the days are reviewed infrequently. Moreover, the methodology does not take some important constraints into account, such as limitations on store shelf space and costs associated with holding additional overstock product in supermarket backrooms.
Kerslake's model simulates the replenishment process, and emulates the rules followed by the forecasting software. "We developed a general model that could be adapted to each store," he said. This systematic approach to planning delivery days has the potential to cut overall delivery costs by more than 20% for some stores. The project also identified a number of other factors which can affect store operations and, therefore, replenishment costs, such as product re-order points and shelf space allocations. In addition, the project determined that specifying the number of deliveries a store should receive each week is not sufficient; the days of the week must also be specified.
For more information on Chris Kerslake's MLOG thesis "A Method for Analyzing the Delivery Frequency from a Distribution Center to a Retail Store" please contact Becky Allen Schneck at: bschneck@mit.edu.