Order Quantity Optimization
This story is about a Netherlands-based developer and manufacturer of specialized industrial vehicles and environmental loading systems. Though being family-owned, they have grown to operate in nine countries, having over two thousand employees. Their manufacturing process involves the purchasing and handling of tens of thousands of SKU’s.
The client’s procurement team had been focusing their efforts on avoiding stock-outs with cost savings being a secondary consideration. In practice, the team used a certain rule of thumb to define standard order quantities and manually adjusted those based on real-world conditions per time period. This process was imprecise, not scalable, and relied on excessive manual (re)work for the purchasing department.
Another important aspect of the challenge is that the client’s production schedule is known up to 12-18 months in advance, which is both a blessing in efficiency (as it allows the client to do batch production of similar types of vehicles together) and a curse. This is because resultingly, SKU requirements fluctuate throughout the year, which the rule of thumb used did not account for.
The Analytics Manager discovered that quantity discount opportunities (material price reductions up to 90%) were being missed out on because of this. The Data Science Manager began to wonder if the company could save money by adjusting its purchasing practices. With a total purchasing amount of hundreds of millions, even an improvement of 1% would yield significant savings. But how to tackle this challenge? It was clear to him that day in day out optimizing purchasing of tens of thousands of SKU’s is beyond reasonable human capacities.
Given our experience in Sourcing and Supply Chain optimization and having been a trusted partner in Process Mining in the last 10+ years, DataLane was asked to also help tackle this challenge.
While the impetus for the project came from effectively leveraging quantity discounts, it became clear that material costs were only one piece of the puzzle. Other factors that complicate the process are vendor agreements as well as transport, holding and obsolescence costs, and of course, avoiding stock-outs. DataLane’s Order Quantity Optimizer (OQO) turned out to be the perfect solution: a mathematical model that computes optimal order quantities per period, yielding minimal total costs within given business restrictions. The model being flexible, all components of the puzzle were customized to the client’s specifics.
A pilot project analyzed the value add for a small set of 250 SKU’s in only one warehouse (out of many). The OQO was tested on historic data (comparing the optimized order quantities to the actual decisions made).
SKUs in Scope
Cost Saving Reached
The pilot resulted in a strong business case. Even with the limited scope of one warehouse, OQO yields up to 75% cost savings per SKU, with total annual savings in the millions.
For the procurement team, confidence in order quantity proposals and process transparency have increased dramatically while manual work has dropped. Based on the success of this project, OQO will be further operationalized and rolled out for all warehouses.