In 1987 Brando Vitali, then Barilla's director of logistics, had expressed strong feelings about finding an alternative approach to order fulfillment. At that time, he noted, "Both manufacturers and retailers are suffering from thinning margins; we must find a way to take costs out of our distribution channel without compromising service." Vitali was seen as a visionary whose ideas stretched beyond the day-today details of a logistics organization. He envisioned an approach that would radically change the way in which the logistics organization managed product delivery. In early 1988 Vitali explained his plan:
I envision a simple approach: rather than send product to the distributors according to their internal planning processes, we should look at all of the distributors' shipment data and send only what is needed at the stores-no more, no less. The way we operate now it's nearly impossible to anticipate demand swings, so we end up having to hold a lot of inventory and do a lot of scrambling in our manufacturing and distribution operations to meet distributor demand. And even so, the distributors don't seem to do such a great job servicing their retailers. There are stockouts despite their holding a couple of weeks of inventory.
In my opinion, we could improve operations for ourselves and our customers if we were responsible for creating the delivery schedules. We'd be able to ship product only as it is needed, rather than building enormous stocks in both of our facilities. We could try to reduce our own distribution costs, inventory levels, and ultimately our manufacturing costs if we didn't have to respond to the volatile demand patterns of the distributors.
We have always had the mentality that orders were an unchangeable input into our process and therefore that one of the most important capabilities we needed to achieve was flexibility to respond to those inputs. But in reality, demand from the end consumer is the input and I think that we should be able to manage the input filter that produces the orders.
How would this work? Every day each distributor would provide us data on what Barilla products it had shipped out of its warehouse to retailers during the previous day, as well as the current stock level for each Barilla SKU (Stock Keeping Unit) Then we could look at all of the data and make replenishment decisions based on our own forecasts. It would be similar to using point-of-sale data from retailers-we would just be responding to see-through information one step behind the retailer. Ideally, we would use actual retail sell-through data, but that's hard to come by given the structure of our distribution channel and the fact that most grocers in Italy aren't equipped yet with the necessary bar-code scanners and computer linkages.
Of course, it's not quite as simple as that. We need to improve our own forecasting systems so we can make better use of the data that we receive. We'll also need to develop a set of decision rules that we can use to determine what to send after we've made a new forecast.
Vitali's proposal, "Just-In-Time Distribution (JITD)," met with significant resistance within Barilla. The sales and marketing organizations were particularly vocal in their opposition to the plan. A number of sales representatives felt their responsibilities would be diminished if such a program were put in place. A range of concerns was expressed from the bottom to the top of the sales organization. The following remarks were heard from Barilla sales and marketing personnel:
• "Our sales levels would flatten if we put this program in place." • "We run the risk of not being able to adjust our shipments sufficiently quickly to changes in selling patterns or increased promotions." • "It seems to me that a pretty good part of the distribution organization is not yet ready to handle such a sophisticated relationship." • "If space is freed up in our distributors'...