Poster
Client Context
The Coca-Cola North America Operating Unit (NAOU) aims to improve its chilled beverage warehousing network with a more systematic approach. Their current reactive strategy involves opening new warehouses in a fragmented manner through third-party logistic providers (3PLs), which leads to increased operational costs, complexity, and utilization issues. The team scope for this project involved 12 plants, 28 warehouses, 948 customers, and served 340 SKU stocks. Our team aims to formulate a long-term warehousing strategy encompassing warehouse placements and capacities, customer assignments, and safety stock quantities. The main objective is to minimize transportation and inventory costs.
Project Objective
We identified two valuable opportunities to reduce the inventory and transportation costs, and one opportunity to aid Coca Cola in future growth. Currently, Coca Cola determined their safety stocks by using the management team’s experiences, thus drives an excess in safety stocks as they want to fulfill their service levels which cause the current warehouses inventory to hold over 42.6% in safety stock. The first opportunity that our team found was to apply the King’s Method to reduce the current inventory, which provided Coca Cola with a mathematical approach in determining their safety stocks and reduced excessive unused inventory. The second opportunity focused on maximizing the warehouses utilization by reassigning plants to warehouses and warehouses to customers. Specifically, considering changing the flow of overutilized to underutilized warehouses. Reassignments can create a more intense pooling effect of demand variability, thus potentially lowering transportation costs and mitigating issues of high warehouse utilization. The final opportunity involves the investigation of potential capacity expansions at current 3PL warehouses and determining the optimal locations for new overflow warehouses prepares Coca-Cola for long-term feasibility with demand growth. Since Coca-Cola has short-term contracts with 3PLs for all but two overflow warehouses, and hence there is an opportunity to re-evaluate the network as these contracts expire. A proactive approach is beneficial because chilled warehousing is at a premium. Addressing these opportunities will helps Coca Cola move closer to their objectives by reducing inventory cost, transportation costs and provided Coca Cola a future plan.
Design Strategy
Our approach to the project was to develop models that were built upon each other, one major system at a time. First, we developed code that calculated a baseline for the current system’s transportation and inventory costs and validated our outputs with our client. Next, we built an optimization model that kept the current system’s network assignments but calculated the optimal amount of safety stock to hold at each warehouse-sku pair. Then we built a model that kept the current system safety stock numbers but optimized the transportation costs by reassigning plant-warehouse-customer pairings. After that we combined the two models together since we knew they were working independently. Finally, we built additional functionality to make current overflow warehouses have a variable amount of capacity and fed in 28 warehouses with estimated inventory and transportation costs that the model could choose to open any of. At every step during development of the optimization models, the outputs were shared with the client in order to capture as many real-world constraints as possible, and ensure our numbers made sense.
Deliverables
Deliverables include a custom dashboard showing detailed recommendations by SKU, customer, warehouse, and plant. It includes safety stock quantities, assignments from plants to warehouses to customers, and maps to visualize the recommended system flows. Additionally, the models used to find these suggestions are provided to help the client dynamically adjust their network as needs arise. Finally, multiple alternative scenarios are provided since modeling cannot capture all the complexities of the real system; parameters and constraints of the model are modified to demonstrate the sensitivity of the outputs.
Value and Impact
By analyzing customer demand variances and recalculating safety stocks by accounting for pooling variability and the desired service target, the team recommends a 56% reduction in the quantity of safety stock held annually for a saving of $6.8M. Furthermore, by reassigning case flows, the team has found potential transportation savings of $19.5M (7.6%).
Additionally, the team recommends a long-term strategy shift from opening spall warehouses closely located to current warehouses and plants in the system to large warehouses that are centrally located with low inventory cost per pallet.
Project Information
Student Team
Caleb Becker, Ariana Garbers, Rakshanda Khan, Pablo Martin Jimenez, Tram Anh Nguyen, Thao Phan, Thi Tran, Rachel Wewengkang
Faculty Advisor
Faculty Evaluator
