Concepts, Strategies, and Case Studies THIRD EDITION David Simchi-Levi Massachusotts Institute of Technology (MIT), Cambridge, Massachusetts Philip Kaminsky University of California, Berkeley Edith Simchi-Levi LogicTools, Inc., Lexington, Massachusetts fay McGraw-Hill (soa Gi Irwin Boston Burr Ridge, IL Dubuque, IA Madison, WI New York San Francisco St, Louis. Bangkok Bogoté Caracas Kuala Lumpur Lisbon London Madrid Mexico City Milan Montreal New Delhi Santiago Seoul Singapore Sydney Taipei TorontoDE Inventory Management and Risk Pooling E Steel Works, Inc. Gry Lemming sat in his new corer office and ‘upped his pencil on the desk. Lemming had just been iumed head of Steel Works, Inc.’s new centralized logistics group. After a decade of experience imple- menting MRP (Materials Requirements Planning) systems throughout the company’s manufacturing tivilities, Lemming was confident he could handle job. Until this morning, ‘Our inventory levels are ridiculous!” barked Jean. ‘Ou Blane, the company's Chief Financial Officer. “Our customer service is the worst in the industry, aund getting worse,” grumbled Kirk Callow, the CEO. J emming started to explain, “You see, I've already set up a team to look at all of that...” But before: he could finish, Callow stood up. “Sales are down 40 percent and expenses are up 25 percent. Our best customers are calling me and telling me they're joing to our competitors, and at the rate we're losing market share we won't be in business in a year. | don't want to hear about teams; I want you back in hore in a week telling me how you're going to fix this thing” Lemming looked over the list of people he"d asked to meet with him this week. He shook his thead—how do I lower expenses and improve per- formance? How will I ever find the right answer? BACKGROUND Steel Works, Inc., is a manufacturer of custom and specialty use steels with annual sales of $400 million in 1993. Founded in 1980 by three brilliant material scientists from MIT, the company now employs more than 2500 people at § different locations. With its first product, DuraBend™, the company earned a reputation as a high technology provider and quickly established @ niche position in what is typically regarded as a commodity market. Its two divisions, Specialty Products and Custom Products, are very separate and distinct businesses. CUSTOM PRODUCTS Lemming’s first interview of the morning was with Stephanie Williams, President of the Custom division. “Our motto is ‘The Customer Comes First, Second, and Third, But Never Last,”” explained Ms. Williams. “The Custom division develops most of its products under contract for a single customer, for sale exclusively to that customer, and works very closely with them from before a product is invented Source: This case was prepared by research assistant David Kletier under the direction of Professor Stephen C. Graves asthe basis for clas discussion rather than to ilasrate either effective or ineffective handling of an aiministratve simation. Copyright © 1996 Massachaseus Institute of Technology. The company, people, data, and events depicted herein are entirely fictitious. Any resemblance {to actual peopl, businesses, or situations is purely coincidental 272B DESIGNING AND MANAGING THE SUPPLY CHAIN Until our product is a part of their product, We have the best scientists and engineers in the world, and that is why the biggest companies in the U.S. come to us. We've designed the metals that make our customers’ products work great. That's why we typically aren’t allowed to sell our products to anyone but the original customer—our customers" ‘competitors would love to buy from us.” Williams went on to explain that eventually when 4 product is no longer leading-edge, the Custom division will negotiate with the customer to allow ‘Steel Works to sell the product to anyone. “Such dis- cussions are an art form,” explains Stephanie, “bu can make a huge difference in sales revenues for us. “Take DuraFlex™ R23, for example. We developed that under contract for one of the big three auto companies, It took us over a year to develop, and there is still no product like it in the marketplace. Yet we were able to convince our customer to allow us to sell it openly on the market at a 30 percent premium over ‘what we charge them. We still sellin large volumes to our customer, and Specialty Products makes a small fortune manufacturing the exact same steel and selling itatahigher price to four other auto manufacturers and ‘copier company.” Williams displayed a schematic of Custom Products" manufacturing system, The three manu- facturing sites were each located within a few miles of one of Custom Products’ three R&D centers, which served the West, Midwest, and Eastern regions of the U.S. Customers and their products were each assigned to a specific plant and R&D center. Steel Works operated several warehouses located near the plants. ‘The only question on Lemming’s mind was why the inventory levels were so high. The reply was direct and blunt; “We've got to keep our customers happy. Customers aren’t satisfied when you tell them that they have to wait three weeks for delivery! We listened to that corporate inventory reduction mandate in 1991 and cut our inventories back 20 percent and ‘we were running out of product every week!” SPECIALTY PRODUCTS “Let me tell you something,” Barry White said as he stormed into the room, “we are nothing like Custom.” Mr. White was President of the Specialty division, whose sales have been the most hard hit in recent months. “That Custom division has nothing to do all day bat play in laboratories. We're the ones out in the marketplace selling every day and bringing in 67 percent of this company's revenue, I've got the best sales force around, and they are what makes this business work.” “Custom thinks they're so special because they've got some big customers; well guess what, so do we. Our largest customer in Specialty brings in 10 percent of the revenue for this company, and it is with blood, sweat and tears that we keep them and everyone else as our customer, You want to salve some problems? Manufacturing is where the problems are; you should talk to them. I've got my plant ‘managers screaming at me every day that the CSR's [customer service representatives] are screaming at them because the customers are screaming at the CSR's for not having any steel in the warehouse to ship, And that’s not the CSR’s fault, it’s manu- facturing’s fault.” “Last week the IS department comes knocking on my door telling me how great it would be if all of Steel Works was on a common computer system, and. wants me to pay $12mm for my division, They think they understand our business but they don’t. We don’t need centralized computer systems, we need to fix manufacturing!” White explained that like the Custom Products division, Specialty attempted to manufacture its products in a single plant. The division operated three plants that manufactured 6 different product lines. The division's general strategy was to exploit economies of scale in production and to rely on the logistics network to distribute the product nationwide. ‘To achieve further efficiency, product families were almost always manufactured in the same plant to ‘save manufacturing costs: the change-over costs ‘between products in the same family were often con- siderably lower than across different families. Products were produced in a rotating sequence. For example, DuraFlex™ R23 js always produced during the first week of the month. And before Lemming knew it, White had stormed out of the room. ANALYSIS It was now Tuesday and 20 percent of the week was gone. Debby Klein, a senior logistics analyst, sat across from Lemming,Wig customers (> $25mm) ‘Small customers (< Simm) All others: ‘Total QUE 2-1 Profile of Specialty Products customers. “Well it’s just like you said it would be. Custom ly 0 lot of products, and something like 90 percent ‘of them are sold to only one customer, On the other Iwacl, Specialty has something like 130 customers wr some 120 products. They've got so many moxtucts I can’t even keep track!” Vichby then related the grim news about customer service levels, “Based on data collected by our order nity systems, approximately 70 percent of the orders rntcred into the system are scheduled to be shipped Nini stock within 48 hours. The rest of the orders {W) percent) are either canceled by the customer at the tin of entry or placed in a backorder file. I couldn't {in out how many of the backorders are canceled, and 1 wasn’t sure if we needed to know that” lemming then asked about the big customer. “Yep, they're big all right. They're like 15 percent of the Specialty sales for 1993 but they buy a lot of «lifferent products. There are other big customers, ‘lnough. And small ones, and medium-sized ones too” ¢ Vigure 2-1) “Thanks Debby,” said Lemming, Jweling more confused than ever. ‘Atter lunch, Lemming had the production plan for Specialty’s Ohio plant faxed to him. The Ohio plant ‘wunufactured the DuraBend™ and DuraFlex™ jiluct families. Production at the plant followed a epular rotating schedule, producing each family ‘hots once per month. The plan seemed consistent with Barry White's account of the division's manu- Jicturing strategy. At the end of the day, a young forecasting expert ‘umned Maria stopped by, looking quite upset. “I waked at all the products like you asked. It's a mess lust like you said, 80 percent of the products fall in this “highly volatile’ category (see Figure 2-2). With .tnndard deviations that large, I don't think a demand {onecasting tool is going to help you very much.” CONSULTING Wright and early on Wednesday morning, Fred Chow, a logistics consultant, walks into Lemming's ulice, “From what you described on the phone, the (GHAPTER 2 INVENTORY MANAGEMENT AND RISK POOLING 29 ‘Specialty Products 20% of Products ev <05 ‘te maar content of aon of ne mons demand eu {othe standard deviation of demand over one mont, vided By ‘ean monty demand. FIGURE 2-2 Demand variation in Specialty Products, answers are all very simple. There are three things you need to do: 1, Get rid of all those products. You've probably got products that have annual sales of a few thousand dollars, and probably have products that aren't selling at all. Discontinue them and focus on your high margin, high volume products to maximize ‘your revenue, 2. Use a statistical forecasting package to predict your ‘demand and this will lower the amount of inventory ‘you need. You see, the inventory levels you need to hold will be a function of the least squares regression and the resulting standard deviation of error in demand in the leadtime. So reduce that and you've reduced your inventory. Violé! 3. You've probably got too many warehouses. Everybody knows that fewer warehouses mean less inventory.” Lemming was now, excited. Although he didn't understand about the least squares-thing, and although Maria said yesterday that forecasting wouldn't work, now he was getting somewhere. Accidentally calling the consultant “Jonah” at one point, Lemming was forever grateful. REALITY SETS IN The businesses completely rejected the idea of dis- continuing the slow-moving products. “We can’t do that! Our most important customers buy those products!” So much for that idea.30 DESIGNING AND MANAGING THE SUPPLY CHAIN, If that weren’t bad enough, Debby happens to stop back in. “Reduce our warehouses? What are you talking about? If we have to ship from fewer warehouses, it will take longer, cost more and that Will really make the divisions mad. Plus, just because you combine two warehouses doesn’t mean you're going to save that much money, Some fixed costs certainly, but that won't make up for those added transportation costs you're going to have to swallow.” Lemming didn't believe this, so he mounted an effort to get to the bottom of the warehouse consoli- dation issue, Several hours and several hundred sheets of paper later, Lemming realized that he had underes- timated the data collection and number-crunching involved with this type of analysis. Although the idea had the potential to reduce costs, his team did not have enough time to look at it this week. ‘THE CLOCK CHIMES ELEVEN It’s Thursday, close to midnight, and Lemming is now sweating, Debby, Maria, and John Thompson, a recent Sloan graduate, are all gathered in the corner office. ‘They've listed a dozen ideas on the blackboard: ABC. analysis, customer segmentations, EOQs, and many more . . . and crossed them all off. Lemming can see Callow's angry face now—his career is slipping away, Now what do you do, John Thompson’? OVERVIEW OF STEEL WORKS, INC,, DATA Since it would be nearly impossible to analyze all of Steel Works’ data, a representative sample of products from 1994 has becn provided to you, Specific data has ‘been provided for portions of two of the product lines: DuraBend™ and DuraFlex™, Five spreadsheets are available CD-ROM to assist you in your analysis: $0121958.XLS Sales of DuraBend™ R12 for cach month and for each customer. MONTHVOL.XLS Total sales of DuraBend™ and DuraFlex™ for each month. PRODBALXLS Production batch sizes for DuraBend™ and DuraFlex™, FINCLDATXLS Unit costs and 1994 selling prices for DuraBend™ and DuraFlex™, EOMINV.XLS DuraBend™ and DuraFlex™ inventories at the end of each month. Note: In the provided, all data (including dollar figures) are in thousands. When drawing inferences from the data, you can safely assume that the DuraBend™ and DuraFiex™ product lines are representative of the entire Specialty division. the book's By the end of this chapter, you will understand the following issues: + How firms cope with huge variability in customer demand. + What the relationship is between service and inventory levels, + What impact lead time and lead time variability have on inventory levels. + What an effective inventory management policy is. + How buyers and suppliers use supply contracts to improve supply chain e. * What approaches can be used to forecast future demand. 2.1 INTRODUCTION In many industries and supply chains, inventory is one of the dominant costs. In the United States, for example, over a tillion dollars is invested in inventory. For many managers, effective supply chain management is synonymous with reducing inventory levels in the supply chain. Of course, this is a very simplistic view of supply chain management—in fact, the goal of cffective inventory management in the supply chain is to have the correct inventory at the right place at the right time to minimize system costs while satisfying customer service requirements. Unfortunately, managingCHAPTER: INVENTORY MANAGEMENT AND RISK POOLING 31 Inventory in complex supply chains is typically difficult, and inventory-related decisions can have a significant impact on the customer service level and supply chain nywicmwide cost. ‘Ay we discussed in Chapter 1, a typical supply chain consists of suppliers and munufacturers, who convert raw materials into finished products, and distribution centers and warehouses, from which finished products are distributed to customers. Inventory can appear in many places in the supply chain, and in several forms: © Ruw material inventory. * Work-in-process (WIP) inventory. * Vinished product inventory. Nach of these needs its own inventory control mechanism or approach. Unfortunately, tletermining these mechanisms is difficult because efficient production, distribution, and Jnventory control strategies that reduce systemwide costs and improve service levels must tuke into account the interactions of the various levels in the supply chain, Nevertheless, tive benefits of determining these inventory control mechanisms can be enormous. Gonerai Motors (GM) has one of the largest production and distribution networks in the world. In 1984 GIM's distribution network consisted of 20,000 supplier plants, 133 parts plants, 31 assembly pants, and 11,000 dealers. Freight transportation costs were about $4.1 billion with 60 percent for material shipments. In addition, GM inventory was valued at $7.4 billion, of which 70 percent was, WIP and the rest was finished vehicles. GM has implemented a decision tool capable of reducing tho combined corporate cost of inventory and transportation. Indeed, by adjusting shipment sizes {L@., Inventory policy) and routes (L., transportation strategy), costs could be reduced by about 26 percent annually [24). If inventory is typically expensive and difficult to manage, why hold it at all? luventory is held for a variety of reasons, and inventory control mechanisms need to luke these reasons into account. Inventory is held duc to 1. Unexpected changes in customer demand. Customer demand has always been hard to predict, and uncertainty in customer demand has increased in the last few yeurs due to 1, The short life cycle of an increasing number of products. This implies that his- torical data about customer demand may not be available or may be quite limited (see Chapter 1). bh. The presence of many competing products in the marketplace. This proliferation of products makes it increasingly difficult to predict demand for a specific model. Indeed, while it is relatively easy to forecast demand across product groups—that is, to forecast demand for all products competing in the same market—it is much more difficult to estimate demand for individual products. We discuss this in more detail in Section 2.3 and in Chapters 6 and 11 2. ‘The presence in many situations of a significant uncertainty in the quantity and quality of the supply, supplier costs, and delivery times. 4, [ead times. Even if there is no uncertainty in demand or supply, there is a need to hold inventory due to delivery lead times. 4, Economies of scale offered by transportation companies that encourage firms to transport large quantities of items, and therefore hold large inventories. Indeed,32 _ DESIGNING AND MANAGING THE SUPPLY CHAIN many of the transportation providers try to encourage large-size shipments by offering all sorts of discounts to shippers (see Chapter 3). Similarly, incentives provided by manufacturers to distributors and retailers motivate buyers to purchase large quantities during manufacturers’ promotional periods and hence lead to high invemtory levels. Unfortunately, although it is clear why inventory is held, holding the right amount at the right time in the appropriate place is frequently difficult: + In 1993, Dell Computer's stock plunged after the company predicted a loss. Dell acknowledged that the company was sharply off in its forecast of demand, resulting in inventory write-downs [218]. * In 1993, Liz Claiborne experienced an unexpected earnings decline, as a con- sequence of higher-than- anticipated excess inventories (219) + In 1994, IBM struggled with shortages in the ThinkPad line due to ineffective inventory management (220]. * In 2001, Cisco took a $2.25 billion excess inventory charge due to declining sales. ‘These examples raise two critical issues in inventory management and demand fore- casting. Since demand is uncertain in most situations, the demand forecast is critical for determining what to order, and when to order it, But what is the relationship between forecast demand and the optimal order quantity? Should the order quantity be equal to, greater than, or smaller than forecast demand? And, if order quantity is different than forecast demand, by how much? We explore these issues throughout the rest of this chapter. The strategy, approach, or set of techniques used to determine how to manage inventory is known as a firm's inventory policy. To decide on an effective inventory policy, managers have to take many characteristics of the supply chain into account; . First and foremost is customer demand, which may be known in advance or may be random. In the latter case, forecasting tools may be used in situations in which his- torical data are available to estimate the average customer demand, as well as the: amount of variability in customer demand (often measured as the standard deviation), 2. Replenishment lead time, which may be known at the time the order is placed, or may be uncertain. 3. The number of different products being considered. These products compete on. budget or space and hence the inventory policy of one product affects the others, 4, The length of the planning horizon, 5. Costs, including order cost and inventory holding cost. ‘a. Typically, order cost consists of two components: the cost of the product and the transportation cost. The product cost may exhibit ecbnomies of scale; that is, the larger the order quantity, the smaller the per-unit price. b. Inventory holding cost, or inventory carrying cost, consists of i. State taxes, property taxes, and insurance on inventories. ii. Maintenance costs. iii, Obsolescence cost, which derives from the risk that an item will lose some of its value because of changes in the market. iv. Opportunity costs, which represent the return on investment that one would receive had money been invested in something else (e.g., the stock market) instead of inventory.CHAPTER 2 INVENTORY MANAGEMENT AND RISK POOLING 33 8 Service level requirements. In situations where customer demand is uncertain, itis tien impossible to meet eustomer orders 100 percent of the time, so management ‘weeds to specify an accepiable level of service.