How simulation-driven re-slotting and process optimization delivered over $1.1 million in annual savings with a 3.6-month payback
Warehouse simulation and optimization have long been used to improve internal operations across the supply chain. However, most simulation environments rely on statistical distributions and hypothetical scenarios rather than actual warehouse flow data. This case study presents a more effective approach: optimizing a distribution center using a full year of real outbound data, accounting for seasonality, SKU slotting, rack types, and pick paths.
Using Simcad Pro, an interactive on-the-fly simulation platform, the team built a complete digital twin of the facility and used it for both slotting analysis and ongoing operational optimization.
The distribution center consists of three main operational zones: a freezer, a cooler, and a general rack pick area. All picked orders are delivered to a limited-space staging area before being loaded onto trucks for delivery. Received pallets are consolidated and routed to storage locations by product type, with shelves replenished from top to bottom as needed. Picking always occurs from the bottom two shelves — the "strike zone."
All receiving is completed before 10 AM, with pallets unloaded, consolidated, and stored in upper shelves. Bottom shelves are replenished to full capacity during this process. Picking begins at 11 AM daily and consists of two streams: Parcel orders (small shipments via UPS, FedEx, DHL, or USPS) and LTL orders (palletized loads shipped by truck).
The simulation was designed to address three interconnected operational issues:
With only 10 docks available, truck delays, consolidation overhead, and limited staging space made it difficult to consistently complete inbound operations by 11 AM.
The order release process needed improvement to reduce staging area congestion, shorten travel distances, minimize empty resource travel, and maximize the grab factor.
Existing warehouse slot assignments did not account for pick frequency, congestion patterns, or grab factor, creating unnecessary inefficiency across all three zones.
A full digital twin of the warehouse was built in Simcad Pro, directly reflecting the physical layout and operational rules of the facility:
Because seasonality plays a significant role in distribution center performance, the model was validated across a full year of operation. Initial validation used a single week of data — approximately 750,000 transactions — loaded directly from the WMS.
In Phase 1, the model replayed the WMS-generated pick sequences and storage moves using built-in operational constraints. In Phase 2, the model generated its own orders, picks, consolidation, and storage operations from raw inbound and outbound datasets, enabling rapid capacity analysis and re-slotting scenarios with minimal setup.
Model accuracy vs. real-time performance — validated across 4 seasonal periods with less than 0.02% variance
Before proceeding to optimization, detailed spaghetti diagrams, congestion analyses, and heat maps were generated for each pick zone and operation type through the Simcad Pro interface.
The simulation-driven analysis produced measurable improvements across every area of operation — without requiring additional racking or capital equipment.
Efficiency gains in the freezer, cooler, and floor pick areas respectively, achieved through two-path slotting optimization and zone-level re-slotting.
Replenishment frequency reduced across the entire warehouse by re-slotting high-velocity SKUs closer to pick zones.
Staging area congestion reduced by replacing the waving system with simulation-driven order releases timed to individual pick durations.
Average truck dock-to-ship time decreased by sequencing order releases with truck delivery and parcel shipment schedules.
Reorganized staging, dock assignment rules, and staffing allocation ensure all receiving completes on time, with 80% staging utilization.
After more than a year of live use, the model's efficiency predictions remain within 98.92% of actual results.
Based on industry benchmarks for a mid-to-large distribution center processing ~750,000 weekly transactions
Estimates are based on a ~150-employee distribution center with $10.9M annual labor costs. Actual results will vary by facility size, labor rates, and operational profile. Avoided capital expenditure of ~$800,000 (racking and deferred expansion) is excluded from annual figures above.
The simulation model has become an integral part of daily operations. Warehouse managers run the model each day to forecast how the operation will perform, and the slotting optimization tool is run biweekly to identify additional efficiency gains as product mix evolves.