Inventory reduction by multi echelon optimization

We performed a project on inventory reduction for a global market leader in flavors and fragrances. Supply chains in this industry are pretty complex and characterized by volatile and erratic demand, production processes over multiple stages and crop-driven raw materials. The company is vertically integrated and produces ingredients (MTS), that are partly sold to 3rd parties and partly compounded in finished products (mainly MTO). The resulting supply chain has complex intercompany good flows and a high level of dependencies.  


The company started a journey to improve their supply chain planning capabilities. As a part of this improvements an inventory assessment was executed. The company already ran a high-quality stock parameter setting in SAP. EyeOn was required to further optimize inventory settings in their supply chain with >60 locations, 13 echelon levels (BoM and locations) with 200.000 products and about 200 million records. 

The project started with a kick-off and data collection, after which we performed data validation and analysis. We validated and performed analyses of lead time, supply uncertainty and demand analysis, batch sizes, FG – RM relation analysis, and safety stock analysis. We furthermore performed base-case determination, identification of low hanging fruit, modeling of intermitted demand, classification of demand and the replenishment policy per class. Next, we determined network effects of lot sizing, scenario optimization, and set policy and determine the optimal settings. We optimized over echelons (e.g., local- and central stocking). We presented our findings in an improvement roadmap and a set of actionable quick wins.  


The project identified considerable inventory savings. We provided the company with data driven insights in inventory drivers (e.g., lead time analysis) and, indicated quick wins (like unnecessary double stocking and rebalancing opportunities). The improved settings by applying state of the art multi echelon inventory models (MEIO) were implemented using a decision tree logic. We extensively trained the planners in their Center of Excellence in a multiple day workshop to understand MEIO and the forthcoming decision tree. Confirmed savings in the first year are over 20 M€. 

Search for