CS: How do you determine the accuracy of your inventory projections?
Zientarski: We look at a forecasting accuracy metric and an inventory accuracy metric. The system calculates a theoretical inventory based on these stats, and compares it to what the actual inventory is. Then our buyers can make inventory projection determinations. They’ve been doing this work a long time, so they have a good feel for it.
We can look at the inventory rates at the buyer level, warehouse level and vendor level. So we can drill down pretty well to see, for example, what’s causing a problem.
CS: How have you achieved that accuracy, and how do you plan to maintain or even improve it?
Zientarski: We’d like to see vendor collaboration on our inventory projection needs in 2003. Vendors could share in the forecasting duties. We don’t currently do that.
We trade files, and we can give the vendor our product needs a year ahead of actual sales. We’re moving toward a system in which our top vendors can see our data via the Internet and help us to forecast up the supply chain. JDA offers such a collaboration system, and we’re looking at that for 2003.
CS: With 70,000 SKUs, however, won’t that be a huge endeavor?
Zientarski: Not really; 40 of our vendors account for 90 percent of our product sales. We have about 300 vendors in all.
CS: Who creates the inventory forecast?
Zientarski: Our purchasing department. The merchandising department puts the catalog together, and before they start on the next book, they meet with the purchasing department to get input on sales forecasts and other things. We can copy forecasts and seasonal profiles, and then apply them to new items.
CS: What forecast tolerance percentages do you use?
Zientarski: We have a deviation rate of +5 to -3 percent.