Navistar seeks to provide customers with trucks that deliver maximum uptime, meaning they are always available for operation. However, dealers want to minimize their investment in parts inventory, so the analytics challenge is to identify which parts each dealer should carry, in what quantity, and at what time.
Our analytics team developed a new dealer part stocking logic that focused on reducing inventory depth on individual parts and replacing it with breadth. This was done by 1) characterizing and clustering parts based on attributes and sales patterns at each dealer and 2) analyzing the minimum inventory required to satisfy customer demand without depleting the dealer’s stock before the next shipment. To achieve this, the team developed an inventory simulation model that tested historical performance using the new stocking recommendations; this was done for millions of Dealer + Part Number combinations to determine the best solution for each dealer. The resulting model and process recommends new parts stocking much sooner than the original system and optimizes inventory quantities for each dealer’s needs.
Navistar has verified the improved performance achieved through the new parts stocking model developed by the analytics team. Inventory breadth, meaning the number of unique parts on dealers’ shelves, is up 22% since implementation. Meanwhile, Navistar’s fill rate, the ability to stock parts requested by dealers, has remained almost constant. Dealer inventory costs are down $70 million, while uptime has increased by 7% due to having the right parts on dealers’ shelves. Dealers report fewer customer complaints due to part delays, and improved ease of doing business due to rightsized part inventory. Nicole Swiercz, Navistar’s parts program manager, called the initiative “truly a cross-functional team effort, with several people from the uptime team, data analytics group and IT coming together to drive successful implementation.”