Supply chain professionals are often confronted with the challenge of managing highly volatile customer shipments resulting from the bullwhip effect. This volatility leads to supply chain-wide inefficiencies, high operational complexity, low service levels and substantial costs.
Companies commonly try to compensate the risk of stock outs by carrying elevated levels of inventory and by paying premiums for freight to make up for lost time as a contingency plan. However, this strategy comes at additional costs, and treats the symptom rather than addressing the underlying issue of supply chain variability.
The thesis sponsor company, which operates in the consumer goods industry, sought to understand if a more uniform distribution of goods can reduce variability and enable improvements in transportation cost, service level and cash (i.e. reduce working capital tied up in inventory).
We developed a new order policy based on the lean leveling principle to develop consistent, predetermined customer shipments.
Applying best practices
Lean concepts have been applied extensively in the manufacturing domain, while distribution processes have remained a relatively unexplored frontier in lean practice. In the thesis, we aimed to realize lean-based gains by replacing large, infrequent batch deliveries with frequent small shipments. Motivated by the success of recent lean initiatives, the sponsor company was looking for ways to continue improving operations based on lean principles.
The company’s shipment volatility is driven by large numbers of SKUs under management and the frequent use of promotions, which is typical for the industry. Fulfilling the company’s on-time service level target has become increasingly challenging. Furthermore, the oscillating order patterns have made it difficult for the company to plan its distribution processes, a cause of high transportation costs.
The most substantial issue posed by shipment variability concerns inadequate inventory levels for both buyers and sellers. Weeks of consistently over-ordering certain items are often followed by periods of order levels at close to zero for the same item. In addition to the previously outlined costs, operating under the current order systems leads to large working capital requirement for inventory which is not immediately required.
To counteract these dynamics, we developed an order policy that created a consistent minimum shipment level while maintaining the flexibility to cover unexpected demand peaks.
The Future is flat
As advocated by lean theory, we focused our analysis on the top selling 50% of SKUs, which were derived from a SKU segmentation. The new order policy for these SKUs consisted of a fixed component (based on a percentage of the historical average) and a variable component. We simulated order policies under varying degrees of implementation of lean leveling and compared their performance with the actual results in a sensitivity analysis.
Our model enables the sponsor company to create more stable customer shipments by determining the optimal ratio of fixed and variable shipments. The optimal order policy balanced the evaluation criteria (transportation cost, service level, cash) which could enable sustainable gains for both the sponsor company and its customers. Overall, we were able to demonstrate that lean leveling reduces shipment variability and leads to improved operational performance.
We believe that our research can be applied to other companies and industries that seek to diminish the effects of shipment variability. Instead of discussing whether to use predetermined customer shipments, businesses may be better served to fine-tune the number of fixed versus variable shipments.
This article was written by Melissa Botero and Fabian Brenninkmeijer, and was published first on the Supply Chain Management Review website. The article is based on the SCM thesis Reducing Shipment Variability Through Lean Leveling written by the authors and supervised by James B. Rice, Jr., Deputy Director, MIT Center for Transportation & Logistics. For more information on the research please contact James B. Rice, Jr. at: firstname.lastname@example.org