Illustrative Case Study — A 6-point service gain with $5M less inventory
- Client profile: $480M industrial distributor operating a multi-site fulfillment network
- Network: 6 distribution centers and approximately 28,000 active SKUs
- Core challenge: Inventory was increasing while stockouts, expedited freight, and customer-service issues persisted
- Engagement: 10 weeks from network diagnosis through a two-distribution-center pilot
- Primary objective: Improve product availability while reducing excess inventory and working-capital requirements
- Decision at stake: Whether the company needed to carry materially more stock—or make better decisions with the inventory it already owned
- Existing operating systems — ERP, WMS, TMS, purchasing, supplier, and inventory data were brought into a common analytical view rather than replaced.
- Demand and lead-time modeling — Historical demand, supplier reliability, order variability, and replenishment patterns were used to understand where existing planning assumptions no longer reflected reality.
- Inventory segmentation — Products were grouped by demand pattern, margin, customer importance, lead-time variability, and substitutability so different SKUs were not managed with the same rules.
- Scenario simulation — Inventory policies were tested under changing demand, supplier delays, and service targets to show how much stock was truly required under different conditions.
- Exception-management layer — Planners were shown the items and situations that actually needed human attention rather than manually reviewing thousands of SKUs.
- Operating dashboard — Leadership could see service level, working capital, expedited freight, stockouts, and inventory movement together instead of optimizing each metric separately.
The problem
The company had increased inventory repeatedly to protect customer service.
It was not working.
Inventory had reached approximately $55M, but important products were still stocking out. Teams were frequently paying for expedited shipments, planners were moving inventory between facilities at the last minute, and customer-service teams were managing exceptions after they happened.
The instinct was to raise safety stock again.
We tested whether that would solve the problem.
What we found
The company did not primarily have an inventory-volume problem. It had an inventory-placement and decision problem.
Some locations carried months of demand for low-risk products while frequently running short on high-value items. Supplier lead-time assumptions were outdated. Replenishment rules treated very different products similarly. Planners had accumulated manual workarounds that made sense individually but created inconsistent decisions across the network.
Experienced planners knew which suppliers routinely missed promised dates and which customers could not tolerate a shortage. That judgment was valuable—but it was not consistently reflected in the planning system.
What we changed
We segmented products by demand pattern, margin, customer importance, lead-time variability, and substitutability.
We then modeled inventory requirements under different demand and supplier scenarios and redesigned replenishment and transfer rules around the actual variability in the network.
A two-distribution-center pilot focused on:
- Revised safety-stock logic
- Supplier lead-time assumptions
- Inventory allocation
- Interfacility transfers
- Expedite approval rules
- Exception prioritization
The system surfaced the exceptions that required human judgment instead of asking planners to manually inspect everything.
The result
Modeled inventory declined approximately 9%, releasing roughly $5M of working capital.
On-time, in-full performance increased from approximately 89% to 95%.
Expedited freight spend declined approximately 24%, or about $1.5M on an annualized basis.
The important result was not simply lower inventory. The network carried less inventory while serving customers more reliably.
