Know Before You Go: Optimal Inventory Deployment

By Ben YoKell, Sr. Manager, Supply Chain Strategy Practice Thursday, December 1, 2011 In the face of mounting pressure on margins, it can be a...

By Ben YoKell, Sr. Manager, Supply Chain Strategy Practice
Thursday, December 1, 2011

In the face of mounting pressure on margins, it can be a challenge to satisfy supply chain management’s most basic demands of reducing total landed delivery costs and maintaining or improving customer service levels.  Combine this with today’s growing complexity and fickle market swings, and an important reality emerges:  To remain competitive, you must ensure the right tradeoffs are considered, and that they are considered as completely and accurately as possible.   One such tactic to evaluate supply chain tradeoffs and mitigate the impact of cost volatility is inventory deployment optimization – that is, deciding where and how much inventory to position in your network.

Though it sounds straightforward, to be effective, the nuances of inventory positioning must be addressed.  We have found that even the “leading approaches” leave significant money on the table and may even result in a faulty supply chain configuration – one with inflated costs, degraded service, and locked-in inefficiencies from procurement through delivery. So just what are these generally accepted leading approaches and why are they falling short?

In general, an optimal deployment strategy must be able to determine which network locations should stock each item to achieve the lowest possible total cost and maintain the desired fill rate. Seems simple enough, but when you factor the uncertainty in demand, lead time, and supply as you optimally try to balance holding costs, replenishment quantities, transportation options, and storage capacity, the problem is challenging at best.

Most analytical planning approaches – and all existing off-the-shelf software solutions – subdivide the elements of optimal inventory deployment into two problems, each solved separately:  network flow path optimization and inventory policy optimization. Therein lays both the explanation of commonly accepted best-in-class approach as well as its primary shortfall.

The network flow path component is familiar to most:  which locations and lanes minimize total sourcing, production, warehousing, and transportation costs, considering the seemingly infinite number of ways to assign customers to servicing locations, items to stocking locations, and sources to receiving locations? The inventory policy optimization component is less familiar:  how much inventory should be held at each stocking location to satisfy fill rates, minimize exposure to lead time uncertainty, and minimize inventory cost?

As stand-alone planning techniques, each of these are high-ROI initiatives; yet the optimal inventory deployment problem presents a special challenge that goes beyond separate treatment of the two sub-problems. So what’s the issue? Quite simply, inventory planning contains a paradox. To determine the optimal amount of inventory considering all supply chain costs, you must assign flow paths; but you can’t assign optimal flow paths without determining how much inventory you need. The optimal amount of inventory needed at a given stocking location depends on which downstream demands and upstream sources are assigned to that stocking location, and the optimal location and lane assignments depend on how much inventory will be required. Inventory is paradoxically both an input to AND a result of the deployment decision.

Not surprisingly, many simply make a deployment decision without considering inventory optimization. This approach is ample under certain conditions, but is far from “industry best.” In fact, if you look carefully into the methodology of most deployment initiatives, inventory optimization is applied after a network flow path optimization. While this approach can also be effective in certain situations, I would not consider it a “best practice.” Why? The resulting strategy may not be optimal or even operationally-feasible when capacity is limited, often today’s reality.

In the past year, we’ve integrated the separate approaches and distinct tool sets into a single deployment problem with success at several $5B+ organizations. The results have been compelling, but not without a fair amount of effort. With that said, we firmly believe that a breakthrough will come on the software front fairly soon that will redefine “best-in-class” inventory deployment. When it happens, you’ll be sure to hear about it on this blog, where we’re focused on pursuing the bleeding edge of SI&OP optimization.  In the meantime, you can click here to read about other cost mitigating tactics.

In this article