1. What is JITD, describe it shortly. 2. What re the advantages and disadvantages of JITD 3. What problems do you found in barilla case? 4. List the lead time and inventory level in each stage of barilla distribution systems. CASE Barilla SpA (A) unwilling to give up their authority to place orders as they pleased; some were even reluctant to provide the detailed sales data upon which Barilla could make Giorgio Maggiali was becoming increasingly frus- delivery decisions and improve its demand forecasts. trated. As director of logistics for the world's largest Perhaps more disconcerting was the internal resis- pasta producer, Barilla SpA (Societa per Aziont tance from Barilla's own sales and marketing organi- translates as "Society for Stockholders" and is inter-zations, which saw the concept as infeasible or preted as "Inc."), Maggiali was acutely aware of the dangerous, or both. Perhaps it was time to discard the growing burden that demand fluctuations imposed on idea as simply unworkable. If not, how might the company's manufacturing and distribution sys- Maggiali increase the chances that the idea would be tem. Since his appointment in 1988 as Director of accepted? Logistics, he had been trying to make headway on an innovative idea proposed by Brando Vitali, who had served as Barilla's director of logistics before COMPANY BACKGROUND Maggiali. The idea, which Vitali called just-in-time Barilla was founded in 1875 when Pietro Barilla distribution (JITD), was modeled after the popular opened a small shop in Parma, Italy, on Via Vittorio "just-in-time" manufacturing concept. In essence, Emanuele. Adjoining the shop was the small "labora- Vitali proposed that, rather than follow the traditional tory" Pietro used to make the pasta and bread prod- practice of delivering product to Barilla's distributors ucts he sold in his store. Pietro's son Ricardo led the on the basis of whatever orders those distributors company through a significant period of growth and, placed with the company, Barilla's own logistics in the 1940s, passed the company to his own sons, organization would instead specify the appropriate" Pietro and Gianni. Over time Barilla evolved from its delivery quantities-those that would more effec- modest beginnings into a large, vertically integrated tively meet the end consumer's needs yet also would corporation with flour mills, pasta plants, and bakery- distribute the workload on Barilla's manufacturing product factories located throughout Italy. and logistics systems more evenly. For two years Maggiali, a strong supporter of Vitali's proposal, had tried to implement the idea, but Source: Copyright 1994 by the President and Fellows of Harvard now, in the spring of 1990, little progress had been college. This case was written by Janice H. Hammond of Harvard made. It seemed that Barilla's customers were simply Business School 143 144 DESIGNING AND MANAGING THE SUPPLY CHAIN price index 15 415 575 1985 76.9 84.4 932 100.0 99.0 1020 106.5 121.7 1280 In a crowded field of more than 2,000 Italian pasta TABLE 5-1 manufacturers, Pietro and Gianni Barilla differenti- ated their company with a high-quality product sup BARILLA SALES, 1960-1901 ported by innovative marketing programs. Barilla Barilia sales Italian wholesale Year revolutionized the Italian pasta industry's marketing fire in billions practices by creating a strong brand name and image 1960 10.8 for its pasta, selling pasta in a sealed cardboard 1970 1980 344 box with a recognizable color pattern rather than in 456 576 bulk, and investing in large-scale advertising pro- 1982 809 grams. In 1968, to support the double-digit sales 1963 728 growth the company had experienced during the 1984 1.204 1960s, Pietro and Gianni Barilla began construction 1985 of a 25-million-square-meter state-of-the-art pasta 1987 1.634 plant in Pedrignano, a rural town 5 km outside Parma. 1988 1.775 The cost of this massive facility-the largest and 1989 2.068 most technologically advanced pasta plant in the 1990 2.390 world-drove the Barillas deeply into debt. In 1971 in 1900, 1.190 - US$1.00 the brothers sold the company to the U.S. multina- 1900 figures are estimet tional firm W. R. Grace, Inc. Grace brought additional Source Bed an compare and International Financial capital investment and professional management Statistics harbook, Intional Monetary Fund practices to the company and launched an important new Mulino Bianco ("White Mill") line of bakery the south, where Barilla held a smaller share of the products. Throughout the 1970s, facing difficult co-market than in the north but where the market was nomic conditions and new Italian legislation that both larger. In addition, Barilla held a 29 percent share of capped retail pasta prices and increased the cost of the Italian bakery products market. living allowances for employees, Grace struggled to In 1990 Barilla was organized into seven divi- make its Barilla acquisition pay off. In 1979, Grace sions: three pasta divisions (Banilla, Voiello, and sold the company back to Pietro Barilla, who by then Brabanti), the Bakery Products Division (manufac- had secured the necessary funds to purchase it. turing medium-to long-shelf-life bakery products). The capital investments and organizational the Fresh Bread Division (manufacturing very short- changes that Grace had brought to Barilla, combined shelf-life bakery products), the Catering Division with improving market conditions, helped Pietro (distributing cakes and frozen croissants to bars and Barilla launch a successful return to the company. pastry shops), and the International Division. Barilla's During the 1980s, Barilla enjoyed an annual growth corporate headquarters were located adjacent to the rate of over 21 percent (see Table 5-1). Growth was Podrignano pasta plant. realized through the expansion of existing businesses, both in Italy and other European countries, as well as through acquisition of new, related businesses. INDUSTRY BACKGROUND By 1990 Barilla had become the largest pasta The origins of pasta are unknown. Some believe it manufacturer in the world, making 35 percent of the originated in China and was first brought to Italy by pasta sold in Italy and 22 percent of the pasta sold in Marco Polo in the 13th century. Others claim that Europe. Barilla's share in Italy comprised its three pasta's origins were rooted in Italy, citing as proof a brands: The traditional Barilla brand represented bas relief on a third-century tomb located near Rome 32 percent of the market; the remaining 3 percent of that depicts a pasta roller and cutter. "Regardless of market share was divided between its Voiello brand its origins," Barilla marketing literature pronounced, (a traditional Neapolitan pasta competing in the high- "Since time immemorial, Italians have adored pasta." priced segment of the semolina pasta market) and its Per capita pasta consumption in Italy averaged Braibanti brand (a high-quality, traditional Parmesan nearly 18 kilos per year, greatly exceeding that of pasta made from eggs and semolina). About half of other western European countries (see Table 5-2). Barilla's pasta was sold in northern Italy and half in There was limited seasonality in pasta demand-for HAPTERS. THE VALUE OF INFORMATION 145 TABLE 5-2 PER CAPITA CONSUMPTION OF PASTA AND BAKERY PRODUCTS, IN KILOGRAMS, 1080 Country Bread Breakfast cereals Pasta Helgium 85.5 10 1.7 mark 29.9 3.7 1.6 5.5 France 68.8 0.6 5.9 6.5 Clermany (West) 61.3 0.7 52 3.1 Croce 70.0 8.0 Ireland 58.4 7.7 17.9 130.9 0.2 17.8 5.9 Notherlands 60.5 1.0 28 Purtugal 70.0 5.7 Spain 873 0.3 28 52 United Kingdom 43.6 7.0 3.6 13.0 Average 70.3 2.5 52 Aripted from European Marketing Dute and Statistics 1992. Euromonitor Plc 19.323 tily Paste example, special pasta types were used for pasta sal- Pasta Manutacturing ads in the summer while egg pasta and lasagna were the pasta-making process is similar to the process by very popular for Easter meals. In the late 1980s the Italian pasta market as a water and for some products, eggs and/or spinach which paper is made. In Barilla plants, flour and whole was relatively flat, growing less than I per meal) were mixed to form dough, which was then cent per year. By 1990 the Italian pasta market was estimated at 3.5 trillion lire. Semolina pasta and fresh pasta were the only growth segments of the TABLE 5-3 Italian pasta market. In contrast, the export market BARILLA PLANT LOCATIONS AND PRODUCTS was experiencing record growth: pasta exports MANUFACTURED, 1980 from Italy to other European countries were Index Plant location Products expected to rise as much as 20 to 25 percent per 1 year in the early 1990s. Barilla's management esti- Brabant Pasta 2 Cagliari mated that two-thirds of this increase would be 3 Foggia Pasta attributed to the new flow of exported pasta to east- 4 Matera Pasta crn European countries seeking low-priced basic 5 Perignano Pasta noodles, biscuits food products. Barilla managers viewed the eastern 6 Viale Barilla Tortelini, noodies, fresh pasta 7 Huropean market as an excellent export opportu- Caserta Pasta, ruke, breadsticks 8 Grissin Bon Breadstida nity, with the potential to encompass a full range of Rubbiano Rusk, breaks pasta products. 10 Milano Panettone, cakes, croissants Pomezia Croissants 12 Mantova Bocuses PLANT NETWORK 13 Molt Snacks 14 Barilla owned and operated an extensive network of Ascoli Snacks, sliced oats 15 Rodo Sauces plants located throughout Italy (see Table 5-3 and 16 Altamus Pour mill Figure 5-1), including large flour mills, pasta 17 Castellano Pour plants, and fresh bread plants, as well as plants pro- 18 Ferrara Flour mil 19 ducing specialty products such as panettone Matera Pour 20 Termo Flour (Christmas cake) and croissants. Barilla maintained 21 Miano Fresh bread state-of-the-art research and development (R&D) 22 Miano Fresh bread facilities and a pilot production plant in Pedrignano 23 Altopascio Fresh bread for developing and testing new products and pro- 24 Padova Fresh bread 25 Torino Fresh bread duction processes. 146 DESIGNING AND MANAGING THE SUPPLY CHAIN 25 17 14 20 16 11 7. Southern CDC 13 14. 19 for example, whether it was made with or without eggs or spinach and whether it was sold as dry or fresh pasta. All of Barilla's non-egg pasta was made 1,5,6,8,9,15 with flour ground from grano duro (high-protein Company headquarters "hard" wheat), the highest-quality flour for making 18 Northern CDC traditional pasta products. Semolina, for example, is 10, 21, 22 a finely ground durum wheat flour. Barilla used 23 flours made from grano tenero (tender wheat), such as farina, for more delicate products such as egg pasta and bakery products. Barilla's flour mills ground flour from both types of wheat. Even within the same family of pasta products, individual products were assigned to plants based on the size and shape of the pasta. "Short" pasta prod- ucts, such as macaroni or fusilli, and long" prod. ucts, such as spaghetti or capellini, were made in CDC separate facilities because of the different sizes of Company headquarters Refers to Table 5.3 index equipment required. FIGURE 5-1 Map of Barilla plant locations and prod- ucts manufactured. CHANNELS OF DISTRIBUTION rolled into a long, thin continuous sheet by sequential Barilla divided its entire product line into two gen- After being rolled to the desired thickness, the dough Fresh" products, including fresh pasta products, sheet was forced through a bronze extruding die screen: the die's design gave the pasta its distinctive which had 21-day shelf lives, and fresh bread, shape. After passing through the extruder, Barilla which had a one-day shelf life. workers cut the pasta to a specified length. The cut "Dry" products, including dry pasta, and longer pieces were then hung over dowels (or placed onto shelf-life bakery products such as cookies, bis- trays) and moved slowly through a long tunnel kiln cuits, flour, bread sticks, and dry toasts. Dry prod- that snaked across the factory floor. The temperature icts represented about 75 percent of Barilla's sales and humidity in the kiln were precisely specified for and had either "long" shelf lives of 18 to 24 each size and shape of pasta and had to be tightly months (e.g. pasta and dried toasts) or "medium" controlled to ensure a high-quality product. To keep shelf lives of 10 to 12 weeks (e.g. cookies). In changeover costs low and product quality high, total, Barilla's "dry products were offered in Barilla followed a carefully chosen production some 800 different packaged stockkeeping units sequence that minimized the incremental changes in (SKUS). Pasta was made in 200 different shapes kiln temperature and humidity between pasta shapes. and sizes and was offered in more than 470 differ After completing the four-hour drying process, the ent packaged SKUs. The most popular pasta prod- pasta was weighed and packaged. ucts were offered in a variety of packaging At Barilla, raw ingredients were transformed into options for example, at any one time Barilla's 5 packaged pasta on fully automated 120-meter-long spaghetti might be offered in a 5-kg package, a production lines. In the Pedrignano plant, the largest 2 kg package, a 1-kg package with a northern and most technologically advanced of Barilla's Italian motif, a 1-kg package with a southern plants, 11 lines produced 9,000 quintals (900,000 Italian motif, a 0.5-kg"northern-motif" package, a kilos) of pasta each day. Barilla employees used 0.5-kg "southern-motif package, a display pallet bicycles to travel within this enormous facility. and a special promotional package with a free bot Barilla's pasta plants were specialized by the type tle of Barilla pasta sauce. of pasta produced in the plant. The primary distinc Most Barilla products were shipped from the tions were based on the composition of the pasta- plants in which they were made to one of two Barilla CHAPTER : THE VALUE OF INFORMATON 147 contral distribution centers (CDCs): the northern Barilla products were distributed through three CDC in Pedrignano or the southern CDC on the out- types of retail outlets: small independent grocers. skirts of Naples. See Figure 5-2. (Certain products, supermarket chains, and independent supermarkets. such as fresh bread, did not flow through the CDC.) In sum, Barilla estimated that its products were Other fresh products were moved quickly through offered in 100,000 retail outlets in Italy alone. the distribution system; fresh product inventory was typically held only three days in each of the CDC. 1. Small independent shops. Small shops were In contrast, cach CDC held about a month's worth of more prevalent in Italy than in other western dry product inventory European countries. Through the late 1980s the Barilla maintained separate distribution systems for Italian government had supported small grocers its fresh and dry products because of their differences (often referred to as "Signora Maria" shops) by in perishability and retail service requirements. Fresh restricting the number of licenses provided to products were purchased from the two CDCs by inde- operate large supermarkets. However, in the early pendent agents (concessionari) who then channeled 1990s the number of supermarkets began to grow the products through 70 regional warehouses located as governmental restrictions abated. throughout Italy. Nearly two thirds of Barilla's dry Approximately 35 percent of Barilla's dry products were destined for supermarkets; these prod- products (30 percent in the north of Italy and ucts were shipped first to one of Barilla's CDCS. 40 percent in the south) were distributed from where they were purchased by distributors. The dis- Barilla's internally owned regional warehouses tributors in turn shipped the product to supermarkets. to small independent shops, which typically Brando Vitali's JITD proposal focused solely on dry held over two weeks of inventory at the store products sold through distributors. The remainder of level. Small shop owners purchased products the dry products was distributed through 18 small through brokers who dealt with Barilla purchas- Barilla-owned warehouses, mostly to small shops ing and distribution personnel. Barilla dry product factories 65% TL Barilla CDC 10% 35% TL TL 90% TI 18 Barilla run depots LIL Grande distribuzione Distributione organizata LIL LIL Chain supermarkets Independent supermarkets Signora Maria Shops TL Delivery in truckload quantities LTL - Delivery in less-than-truckdood quantities Note: Shipping percentages are based on product weight FIGURE 5-2 Barilla distribution patterns. 148 DESIGNING AND MANAGING THE SUPPLY CHAIN 2. Supermarkets. The remaining dry products SALES AND MARKETING were distributed through outside distributors to Barilla enjoyed a strong brand image in Italy. Its supermarkets 70 percent to supermarket chains marketing and sales strategy was based upon a com and 30 percent to independent supermarkets. A bination of advertising and promotions. supermarket typically held from 10 to 12 days of dry product inventory within the stores, and on Advertising average carried a total of 4,800 dry product Barilla brands were heavily advertised. Advertising products in multiple types of packages, most copy differentiated Barilla pasta from basic com- retailers would carry a product in only one (and modity "noodles" by positioning the brand as the at most two) packaging options. highest quality, most sophisticated pasta product Dry products destined for a supermarket "Barilla: a great collection of premium Italian pasta." available. One ad campaign was built on the phrase chain were distributed through the chain's own distribution organization, known as a grande The collection" dimension was illustrated by show- distribuzione ("grand distributor"), or GD those ing individual uncooked pasta shapes against a black destined for independent supermarkets were background, as though they were jewels, evoking a channeled through a different set of distributors sense of luxury and sophistication. Unlike other known as distribuzione organizzata ("organized pasta manufacturers, Barilla avoided images of tradi- distributors"), or DOS. A DO acted as a central tional Italian folklore, preferring modern, sophisti- buying organization for a large number of inde cated settings in major Italian cities. pendent supermarkets. Most DOs had regional Advertising themes were supported by sponsor- operations, and the retailers they served usually example, Barilla engaged tennis stars Steffi Graf to ships of well-known athletes and celebrities. For got their products from only a single DO. Due to regional preferences and differences in promote Barilla products in Germany and Stefan retail requirements, a typical distributor might dis- Edberg in the Scandinavian countries. Luminaries Most distributors handled products coming from focused on developing and strengthening loyal rela- tribute iso of Barilla's 800 dry product SKUs such as actor Paul Newman were also used to promote about 200 different suppliers; of these, Barilla typ- ically would be the largest in terms of the physical tionships with Italian families by using messages such volume of products purchased. Distributors typi- as "Where there is Barilla, there is a home." cally carried from 7,000 to 10,000 SKUs in total. Trade Promotion However, their strategies varied. For example, one of Barilla's largest Dos, Cortese, carried only 100 Barilla's sales strategy relied on the use of trade pro of Barilla's dry products and carried a total of only motions to push its products into the grocery distri- 5,000 SKUS bution network. A Barilla sales executive explained Both GDs and Dos purchased products from the logic of the promotion-based strategy the Barilla CDCs, maintained inventory in their We sell to a very old-fashioned distribution system. The own warehouses, and then filled supermarket buyers expect frequent trade promotions, which they then orders from their warehouse inventory. A distribu- pass along to their ows customers. So a store will know tor's warehouse typically held a two-week supply right away if another store is buying Barilla pasta at a dis of Barilla dry products in inventory. count. You have to understand how important pasta is in Many supermarkets placed daily orders with Italy. Everyone knows the price of pasta. If a store is sell distributors; the store manager would walk up and ing past at a discount one week consumers notice the down the store aisles and note cach product that reduced price immediately needed to be replenished and the number of boxes Barilia divided each year into 10 or 12 "canvass required (more sophisticated retailers used hand- periods, typically four to five weeks in length, each held computers to record order quantities as they corresponding to a promotional program. During checked the shelves). The order would then be any canvass period, a Barilla distributor could buy transmitted to the store's distributor, orders were as many products as desired to meet current and typically received at the store 24 to 48 hours after future needs. Incentives for Barilla sales representa- receipt of the order at the distribution center tives were based on achieving the sales targets set CHAPTERS THE VALUE OF INFORMATION 149 for each canvass period. Different product cate to be delivered from the following Wednesday wories were offered during different canvass periods, through the following Tuesday. Distributors' sales with the discount depending on the margin structure volumes varied; small distributors might order only of the category. Typical promotional discounts were one truckload a week, whereas the largest warranted 14 percent for semolina pasta, 4 percent for egg deliveries of as many as five truckloads a week pasta, 4 percent for biscuits, 8 percent for sauces, Most distributors used simple periodic review and 10 percent for breadsticks. inventory systems. For example, a distributor might Barilla also offered volume discounts. For exam- review inventory levels of Barilla products each plo, Barilla paid for transportation, thus providing Tuesday, the distributor would then place orders for incentives of 2 to 3 percent for orders in full truckload those products whose levels fell below the reorder quantities. In addition, a sales representative might level. Nearly all of the distributors had computer- Infor a buyer a 1.000 lire per carton discount (repre- supported ordering systems, but few had forecasting Nenting a 4 percent discount) if the buyer purchased a systems or sophisticated analytical tools for deter- minimum of three truckloads of Barilla egg pasta. mining order quantities. Sales Representatives Impetus for the JITD Program Barilla sales representatives serving Dos spent an As the 1980s progressed, Barilla increasingly felt estimated 90 percent of their time working at the the effects of fluctuating demand. Orders for Barilla Nore level. In the store, sales reps helped merchan- dry products often swung wildly from week to week Wise Barilla products and set up in-store promotions; (see Figure 5-3). Such extreme demand variability took note of competitive information, including strained Barilla's manufacturing and logistics opera competitors' prices, stockouts, and new product tions. For example, the specific sequence of pasta introductions, and discussed Barilla products and production necessitated by the tight heat and humid- urdering strategies with store management . In addi- ity specifications in the tunnel kiln made it difficult tion, each sales rep spent a half day in a regularly to quickly produce a particular pasta that had been scheduled weekly meeting with the distributor's sold out owing to unexpectedly high demand. On Huyer, helping the distributor place its weekly order, the other hand, holding sufficient finished goods explaining promotions and discounts, and settling inventories to meet distributors' order requirements problems such as returns and deletions associated was extremely expensive when weekly demand fluc- with the last delivery. Each rep carried a portable tuated so widely and was so unpredictable. computer for inputting distributor orders. The rep Some manufacturing and logistics personnel also would spend a few hours a week at the CDC, favored asking distributors or retailers to carry addi- discussing new products and prices, covering prob- tional inventory to check the fluctuation in distribu- Toms concerning the previous week's deliveries, and tors' orders, noting that with their current inventory settling disputes about different discounts and deal levels, many distributors' service levels to the retailers structures were unacceptable (see Figure 5-4 for sample distribu In contrast, a very small sales force served the GDs. tor inventory levels and stockout rates). Others felt The GD sales force rarely visited GD warchouses and that the distributors and retailers were already carry usually sent their orders to Barilla via fax. ing too much inventory. In the late 1980s a Barilla logistics manager discussed retail inventory pressure DISTRIBUTION Our customers are changing. And do you know why they Distributor Ordering Procedures are changing? As I see it, they are realizing they do not have enough room in their stores and warehouses to carry Most distributors-GDs and Dos aliko-checked the very large inventories manufactures would like them their inventory levels and placed orders with Barilla to. Think of shelf space in retail outlets. You cannot easily once a week. Barilla products would then be shipped increase it. Yet manufacturers are continuously introducing to the distributor over the course of the week that new products, and they want retailers to display each prod- started 8 days after the order was placed and ended 14 uct on the front of their shelves! That would be impossible days after the order was placed-the average lead even if supermarkets were made from rubber time was 10 days. For example, a large distributor that ordered every Tuesday might order several truckloads Claudio Ferrocza. The Pedrignano Worker Mila: GEA, 1988) 150 DESIGNING AND MANAGING THE SUPPLY CHAIN Onders from Cortese Northeast DC to Pedrignano CDC Orders (in quintals) 1000 900 800 700 600 500 400 300 Man 300 quintals Standard = 227 quintals deviation 100 1357 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 FIGURE 5-3 Weekly demand for Barila dry products from Cortese's Northeast Distribution Center to the Pedrignano CDC, 1989. Distributors felt similar pressure to increase the expressed strong feelings about finding an alternative inventory of items they already stocked and to add approach to order fulfillment. At that time, he noted. items they currently did not carry. In 1987 Brando "Both manufacturers and retailers are suffering from Vitali, then Barilla's director of logistics, had thinning margins; we must find a way to take costs Sales and Stockout at Cartese Northeast DC Stockouts 450 400 Sales (quintals per week) 300 250 200 150 100 Stockouts (from DC to retailers Sales 50 05 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 49 51 Week Inventory at Cortese DC 1600 Inventory (in quintals 1200 1000 800 400 200 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 Week FIGURE 5-4 Sample stockout and Inventory levels, Cortese's Northeast Distribution Center, 1989. CHAPTER 6THE VALUE OF INFORMATION 151 out of our distribution channel without compromis- Vitali's proposal, "just-in-time distribution," met ing service." Vitali was seen as a visionary whose with significant resistance within Barilla. The sales ideas stretched beyond the day-today details of a and marketing organizations were particularly vocal logistics organization. He envisioned an approach in their opposition to the plan. A number of sales that would radically change the way in which the representatives felt their responsibilities would be logistics organization managed product delivery. In diminished if such a program were put in place. A curly 1988 Vitali explained his plan: range of concems was expressed from the bottom to I envision a simple approach: rather than send product to remarks were heard from Barilla sales and marketing the top of the sales organization. The following the distributors according to their internal planning pro- cesses, we should look at all of the distributors' shipment personnel: data and send only what is needed at the stores-no more. . "Our sales levels would flatten if we put this pro- no less. The way we operate now it's nearly impossible to anticipate demand swings, so we end up having to hold a . "We run the risk of not being able to adjust our gram in place." lot of inventory and do a lot of scrambling in our manu facturing and distribution operations to meet distributor shipments sufficiently quickly to changes in sell- demand. And even so, the distributors don't seem to do ing pattems or increased promotions." such a great job servicing their retailers. Look at the . "It seems to me that a pretty good part of the distri- stockouts (see Figure 5-4) these DOs have experienced in bution organization is not yet ready to handle such the last year. And that's despite their holding a couple of a sophisticated relationship." weeks of inventory . "If space is freed up in our distributors' warehouses In my opinion, we could improve operations for our when inventories of our own product decrease, we selves and our customers if we were responsible for creat run the risk of giving our competitors more distrib- ing the delivery schedules. We'd be able to ship product utor shelf space. The distributors would then push only as it is needed, rather than building enormous stocks our competitors' product more than our own, since in both of our facilities. We could try to reduce our own distribution costs, inventory levels, and ultimately our . "We increase the risk of having our customers once something is bought it must be sold." manufacturing costs if we didn't have to respond to the volatile demand pattems of the distributors stock out of our product if we have disruption in We have always had the mentality that orders were an our supply process. What if we have a strike or unchangeable input into our process and therefore that one some other disturbance?" of the most important capabilities we needed to achieve . "We wouldn't be able to run trade promotions with was flexibility to respond to those inputs. But in reality, JITD. How can we get the trade to push Barilla demand from the end consumer is the input and I think product to retailers if we don't offer some sort of that we should be able to manage the input filter that pro incentive? duces the orders. . "It's not clear that costs would even be reduced. If How would this work? Every day each distributor would a DO decreases its stock, we at Barilla may have to provide us data on what Barilla products it had shipped out of its warehouse to retailers during the previous day, as well increase our own inventory of those products for as the current stock level for each Barilla SKU. Then we which we cannot change production schedules due could look at all of the data and make replenishment deci to our lack of manufacturing flexibility." sions based on our own forecasts. It would be similar to Vitali countered the concerns of the sales organi- using point-of-sale data from retailerswe would just be zation: responding to see through information one step behind the retailet. Ideally, we would use actual retail sell-through I think JTD should be considered a selling tool, rather data, but that's hard to come by given the structure of our than a threat to sales. We're offering the customer addi- distribution channel and the fact that most grocers in Italy tional service at no extra cost. In addition, the program aren't equipped yet with the necessary bar code scanners will improve Barilla's visibility with the trade and make and computer linkages distributors more dependent on it should improve the Of course, it's not quite as simple as that. We need to relationships between Barilla and the distributors rather improve our own forecasting systems so we can make bet- than harm them. And what's more, the information regard ter use of the data that we receive. We'll also need to ing the supply at the distributors' warehouses provides us develop a set of decision rules that we can use to deter with objective data that would permit us to improve our mise what to send after we've made a new forecast own planning procedures 152 DESIGNING AND MANAGING THE SUPPLY CHAIN Giorgio Maggiali, head of materials management for Barilla's fresh products group, was appointed director of logistics in late 1988 when Vitali was pro- moted to head one of the company's new divisions. Maggiali was a hands-on manager, known for his orientation to action. Shortly after his appointment, Maggiali appointed a recent college graduate, Vincenzo Battistini, to help him develop and imple- ment the JTD program. Maggiali recounted his frustrations in implement- ing the JITD program: In 1988 we developed the basic ideas for the approach we wanted to use and tried to convince several of our distribu- tors to sign on. They weren't even interested in talking about it; the manager of one of our largest distributors pretty much summed up a lot of the responses we had when he cut off a conversation saying, "Managing stock is my job; I don't need you to see my warehouse or my fig- ures. I could improve my inventory and service levels myself if you would deliver my orders more quickly. I'll make you a proposalI'll place the order and you deliver within 36 hours." He didn't understand that we just can't respond to wildly changing orders without more notice than that. Another distributor expressed concerns about becoming too closely linked to Barilla. "We would be giv- ing Barilla the power to push product into our warehouses just so Barilla can reduce its costs." Still another asked, "What makes you think that you could manage my inven- tories any better than I can?" We were finally able to convince a couple of our dis- tributors to have in-depth discussions about the JITD pro- posal. Our first discussion was with Marconi, a large. fairly old-fashioned GD. First Battistini and I visited Marconi's logistics department and presented our plan. We made it clear that we planned to provide them with such good service that they could both decrease their invento- ries and improve their fill rate to their stores. The logistics group thought it sounded great and was interested in con- ducting an experimental run of the program. But as soon as Marconi's buyers heard about it, all hell broke loose. First the buyers started to voice their own concerns; then, after talking to their Barilla sales reps, they started to repeat some of our own sales department's objections as well. Marconi finally agreed to sell us the data we wanted, but otherwise things would continue as before with Marconi making decisions about replenishment quantities and timing. This clearly wasn't the type of relationship we were looking for, so we talked to other distributors, but they weren't much more responsive. We need to regroup now and decide where to go with JITD. Is this type of program feasible in our environment? If so, what kind of customer should we target? And how do we convince them to sign up