Case Study
Computer Assisted Picking Wave Construction

Statement of Problem

Retailers have announced initiatives to lower inventory while always being in stock. This is evidenced by WalMart's announcement to reduce inventory by 20%. In an environment where the manufacturer's distribution infrastructure is already stretched to the limit responding to replenishment reorders, the retailer's objective can only be accomplished by more frequent smaller orders. Coupled with the fact that the Warehouse Management Systems being employed do not have tools to optimize picking waves but instead rely on the analytic ability of the wave manager to construct efficient waves, often distribution costs climb while efficiencies trail off in an otherwise world-class supply chain environment.

Study Setting

The study was conducted analyzing WMS data provided by a manufacturer replenishing multiple "big-box" retailers on a weekly basis. Replenishment orders typically consist of 20,000 to 30,000 units spread across 300 to 400 SKU's. Shipments are packed at the store level for 500 to 1,000 stores. The packing area can accommodate either 60 or 120 stores. The merchandise is picked in batches from a combination of full cartons from high bay storage and individual, (less than full carton), from flow racks. When building waves, the wave manager, where possible, segregates the orders by Strategic Business Unit, (ladies, youth and men's), selects groups of either 60 or 120 stores, as dictated by the constraints of the packing area, into each wave.

Analysis

Case 1: A weekly replenishment order from a major retailer that had been waved manually as described above was "re-waved" using a sophisticated software tool developed by Inner Harbor Solutions that optimized the grouping of SKU's into the same wave to reduce both picks and replenishment of the flow racks while increasing the units per pick for both full cartons and flow rack picks. In this case the average picks per SKU decreased from 2.32 to 1.49 for a reduction of 36%. Flow rack replenishment decreased by 30%.

Case 2: Weekly replenishment orders for 2 customers were combined for waving utilizing the wave planning and optimization tool. The average picks per SKU decreased to 1.35 from 1.49 in case #1. In the current environment utilizing the "separation by store" method, the picks per SKU would have been additive creating in excess of 4.5 picks per SKU versus the reduction achieved by the waving tool. This would result in a picking reduction of approximately 70% in this environment.

Conclusion

The wave planning and optimization tool can achieve a picking labor reduction of 30% to 35% without adversely impacting the process in other areas, (such as separation of outbound freight at the end of the picking process). With proper attention given to the processing of packed cartons, picking labor can be reduced in excess of 50% up to 75%.