44

CHAPTER 3

Thus, in a pull-based supply chain, we typically see a significant reduction in system inventory level, enhanced ability to manage resources, and a reduction in system costs when compared with the equivalent push-based system. On the other hand, pull-based systems are often difficult to implement when lead times are so long that it is impractical to react to demand information. Also, in pull-based systems, it is frequently more difficult to take advantage of economies of scale in manufacturing and transportation because systems are not planned far ahead in time. These advantages and disadvantages of push and pull supply chains have led companies to look for a new supply chain strategy that takes advantage of the best of both. Frequently, this is a push-pull supply chain strategy.
3.2.3 Push-Pull Supply Chain

In a push-pull supply chain, some stages of the supply chain, typically the initial stages, are operated in a push-based manner, whereas the remaining stages employ a pull-based strategy. The interface between the push-based stages and the pull-based stages is known as the pushpull boundary. To better understand this strategy, consider the supply chain time line: the time that elapses between procurement of raw material, i.e., the beginning of the time line, and the delivery of an order to the customer, i.e., the end of the time line. The push-pull boundary is located somewhere along the time line and indicates the point in time when the firm switches from managing the supply chain using a push strategy to managing it using a pull strategy. This is illustrated in Figure 3-1. Consider a personal computer (PC) manufacturer who builds to stock and thus makes all production and distribution decisions based on forecast. This is a typical push system. By contrast, an example of a push-pull strategy is one in which the manufacturer builds to order. This implies that component inventory is managed based on forecast but that final assembly is in response to a specific customer request. Thus the push portion of the manufacturers supply chain is that portion prior to assembly, whereas the pull part of the supply chain starts with assembly and is performed based on actual customer demand. The push-pull boundary is at the beginning of assembly. Observe that in this case the manufacturer takes advantage of the fact that aggregate

SUPPLY CHAIN INTEGRATION

45

Push pull boundary

Push strategy Raw Materials

Pull strategy End Customer

Supply chain time line

Figure 3-1 Push-pull supply chains.

forecasts are more accurate (see Chap. 2). Indeed, demand for a component is an aggregation of demand for all finished products that use this component. Since aggregate forecasts are more accurate, uncertainty in component demand is much smaller than uncertainty in finished goods demand, and this leads to safety stock reduction. Dell Computers has used this strategy very effectively and is an excellent example of the impact of the push-pull strategy on supply chain performance. Postponement, or delayed differentiation in product design (see Chap. 7), is also an excellent example of a push-pull strategy. In postponement, the firm designs the product and the manufacturing process so that decisions about which specific products are being manufactured can be delayed as long as possible. The manufacturing process starts by producing a generic or family product, which is differentiated to a specific end product when demand is revealed. The portion of the supply chain prior to product differentiation typically is operated using a push-based strategy. In other words, the generic product is built and transported based on a long-term forecast. Since demand for the generic product is an aggregation of demand for all its corresponding end products, forecasts are more accurate, and thus inventory levels are reduced. In contrast, customer demand for a specific end product typically has a high level of uncertainty, and thus product differentiation occurs only in response to individual demand. Thus the portion of the supply chain starting from the time of differentiation is pull-based.

46

CHAPTER 3

3.2.4 Identifying the Appropriate Supply Chain Strategy

What is the appropriate supply chain strategy for a particular product? Should the firm use a push-based supply chain strategy, a pull-based strategy, or a push-pull strategy? Figure 3-2 provides a framework for matching supply chain strategies with products and industries. The vertical axis provides information on uncertainty in customer demand, whereas the horizontal axis represents the importance of economies of scale, either in production or in distribution. Everything else being equal, higher demand uncertainty leads to a preference for managing the supply chain based on realized demand: a pull strategy. Alternatively, smaller demand uncertainty leads to an interest in managing the supply chain based on a long-term forecast: a push strategy. Similarly, everything else being equal, the higher the importance of economies of scale in reducing cost, the greater is the value of aggregating demand, and thus the greater is the importance of managing the supply chain based on long-term forecast, a push-based strategy. If economies of scale are not important, aggregation does not reduce cost, so a pull-based strategy makes more sense. In Figure 3-2 we partition the region spanned by these two dimensions into four boxes. Box I represents industries (or, more precisely, products) that are characterized by high uncertainty and by situations in which economies of scale in production, assembly, or distribution are not important, such as the computer industry. Our framework suggests that a high degree of pull-based supply chain strategy is appropriate for these industries and products. This is exactly the strategy employed by Dell Computers. Box III represents products that are characterized by low demand uncertainty and important economies of scale. Products in the grocery industry such as beer, pasta, and soup belong to this category. Demand for these products is quite stable, whereas reducing transportation cost by shipping full truckloads is critical for controlling supply chain cost. In this case a pull strategy is not appropriate. Indeed, a traditional push-based retail strategy is appropriate because managing inventory based on long-term forecasts does not increase inventory holding costs, whereas delivery costs are reduced by leveraging economies of scale. Boxes I and III represent situations in which it is relatively easy to identify an efficient supply chain strategy. In the remaining two cases there is a mismatch between the strategies suggested by the two

SUPPLY CHAIN INTEGRATION

47

Demand uncertainty Pull H I Computer IV Books & CDs L Pull II Furniture III Grocery H Push

Push

Economies of scale

Figure 3-2 Matching supply chain strategies with products.

attributes, uncertainty and the importance of economies of scale. Indeed, in these boxes uncertainty pulls the supply chain toward one strategy, whereas economies of scale push the supply chain in a different direction. For instance, box IV represents products characterized by low demand uncertainty, indicating a push-based supply chain, and low economies of scale, suggesting a pull-based supply chain strategy. Many high-volume/fast-moving books and CDs fall in this category. In this case a more careful analysis is required because both traditional retail push strategies and more innovative push-pull strategies may be appropriate, depending on the specific costs and uncertainties. We discuss this choice in more detail in Section 3.4. Finally, box II represents products and industries for which uncertainty in demand is high, whereas economies of scale are important in reducing production and/or delivery costs. The furniture industry is an excellent example of this situation. Indeed, a typical furniture retailer offers a large number of similar products distinguished by shape, color, fabric, and so forth, and as a result, end-item demand uncertainty is very high. Unfortunately, these are bulky products, and hence delivery costs are also high. Thus, in this case, there is a need to distinguish between the production and the distribution strategies. The production strategy has to follow a pull-based strategy because it is impossible to make production decisions based on long-term forecasts. On the other hand, the distribution strategy needs to take advantage of economies of scale in order to reduce transportation cost. This is exactly the strategy

48

CHAPTER 3

employed by many retailers that do not keep any inventory of furniture. When a customer places an order, it is sent to the manufacturer, who orders the fabric and produces to order. Once the product is ready, it is shipped, typically using truckload carriers, together with many other products, to the retail store and from there to the customer. For this purpose, the manufacturer typically has a fixed delivery schedule, and this is used to aggregate all products that are delivered to stores in the same region, thus reducing transportation costs due to economies of scale. Hence the supply chain strategy followed by furniture manufacturers is, in some sense, a pull-push strategy, where production is completed based on realized demand, a pull strategy, and delivery is according to a fixed schedule, a push strategy. The automobile industry is another example of the conditions of box II. A typical car manufacturer offers a large number of similar products distinguished by functionality, motor power, shape, color, number of doors, sports wheels, and so forth, and as a result, demand uncertainty for a particular configuration is very high. Delivery costs are quite high as well. Traditionally, this industry has employed a push-based supply chain strategy, building inventory for the dealer distribution systems. Thus the automobile industry does not currently follow the model developed in Figure 3-2. In 2000, however, General Motors (GM) announced a dramatic vision for restructuring the way it is designing, building, and selling its products.2 The goal is to allow customers to customize and order cars online and have the cars delivered to the customers door in less than 10 days. GM is moving exactly in the direction predicted by our model toward a build-to-order strategy. Unfortunately, lead times in the automobile industry are currently long: 50 to 60 days on average. To achieve its vision, GM has to redesign the entire supply chain, including the way it partners with suppliers, the way it manufactures products, and the way it distributes products. Reducing lead times to 10 days or below also may require a significant reduction in the number of options and configurations offered to buyers.
3.2.5 Implementing a Push-Pull Strategy

The framework developed in the preceding section attempts to characterize the appropriate level of pull and push for different products.