There are marked differences between supply chain designs. The challenge for
supply chain managers is to acknowledge that they may no longer be using the
optimal designs for the requirements of their businesses. It is widely accepted that the design of a product is responsible for more than 80 percent of its lifetime cost. In the same way, the design of a supply chain has a tremendous impact on the cost and value attributes of the product over its lifetime. The impact will only continue to grow as the battleground shifts in the 21st century from competition between organizations to competition between supply chains. Supply chain design will become a key source of competitive advantage. When considering supply chain designs, we can exploit some fundamental principles to enhance product flow across the value stream and to respond quickly to changing customer expectations. 1 However, supply chains are not static. The supply chain manager must continuously fine-tune planning and execution systems and the software that supports them to match evolving industry dynamics. What are these dynamics? The first is the fast-changing business landscape, itself. Customers are now requiring higher levels of service and attention, and the move toward personalization- the so-called "market of one" concept-puts pressure on those supply chains that are geared to mass markets. Second, competitors may deploy supply chains that give them an immediate edge, as happened when Dell launched its direct-sales, configure-to- order business model. Finally, there is now a wider range of more flexible supply chain designs, with fewer barriers to switching from one design to another. Complicating matters, however, is the fact that supply chain managers long familiar with an existing supply chain design may find it hard to understand or embrace different designs that better suit new market conditions. In this article, we review the four main types of supply chain design, examining their attributes and weaknesses. We argue that supply chain managers today must be prepared to review the efficacy of their current supply chain designs-and be ready to alter designs to better fit their companies' business needs.
*Four Types of Supply Chain Design *
Supply chains deliver products to the customer using one of the following four basic process structures.
1. Build-to-Stock (BTS).
The product is built prior to demand with a standard bill of materials-for example, Diet Cola. The BTS supply chain has the fastest response time to the customer. The customer order is placed and satisfied either from a retail shelf or from a finished-goods stocking point. Because the customer values immediate response for a BTS product, "impulse" products, such as many types of consumer goods, are supplied using a BTS model. However, the price of this immediate satisfaction is some loss of selectivity. The customer takes what is available in predetermined configurations supplied by the manufacturer. One common consequence of such limited choice: The
customer may purchase more product features than actually desired. The BTS model is by no means limited to discretionary consumer purchases. Many critical repair components, such as aircraft components, are supplied using a BTS supply chain design.
2. Configure-to- Order (CTO).
The product is assembled to demand with standard modules or components. Desktop computers offer an example. The CTO supply chain introduces orders prior to assembly and pushes the order to the customer but replenishes (pulls) parts to build he order. In this arrangement, the customer receives greater end-item choice but
sacrifices some of the immediacy of order fulfillment. The automobile industry offers another good example. Automakers and their distributors and dealers are in the initial stages of implementing CTO supply chains. The goal is to offer the customer a wider selection of color/option combinations than is typically available on the dealer lot. However, the customer will not be able to drive off the lot at the time of purchase; she must wait until the automobile is assembled to her specifications. A critical issue for those using (or considering) CTO supply chains is how quickly the customer's needs are satisfied; in particular, how much can they reduce the leadtime from assembly to final delivery. The North American automobile industry is now
targeting delivery of a custom-assembled car within a week of the order being placed, compared to the multi-week window in which it operates today.
3. Build-to-Order (BTO).
The product is fabricated and assembled to order with a standard bill of materials. Examples include executive jets and industrial machinery. In the BTO supply chain, customer orders are introduced prior to fabrication or at the start of the production process. BTO products are usually highly customized to customer specification, very costly to manufacture, or both. The BTO planning requirements are captured in a typical materials-requireme nts-planning (MRP) structure. In effect, BTO commits components, sometimes far upstream into the supply base, to specific customer orders. Once requirements are established, the supply chain produces to specific quantities and due dates. This is in contrast to the pull structure of BTS and CTO supply chains that respond to replenishment signals inside a planned capacity stream. Such locked-in quantities and due dates mean that MRP execution is subjected to significant
expediting and exception activities. Any disruption in the flow of materials can cause due-date slippage throughout the complete build sequence. So the typical MRP-driven supply chain reshuffles purchase-order due dates, the dispatch list,
and customer promises. These actions magnify the variations in capacity up and down the supply chain.
4. Engineer-to- Order (ETO).
The product is fabricated and assembled to order with unique parts and drawings. Example: A thermo-chemical reactor or the U.S. space station. This type of supply chain responds to a truly customized product that requires unique drawings and parts. Custom products manufactured for highly specific uses fit well into this category. The leadtime from order to delivery is often long because of the product's custom nature. Indeed, the front-end engineering is often a neglected but costly process within his supply chain. MRP planning prevails in ETO. The ETO supply chain is the prototypical single-lot, job-shop environment. Upstream planning and logistics are often varied and complex compared to downstream distribution. Distribution and transportation of ETO products are often planned in units of one.
*Design Trade-Offs *
Each supply chain alternative presents different value trade-offs for its participants. For example, in the BTS supply chain, suppliers speculate on assured demand by moving goods forward to satisfy the customer's immediate demands. Any forecast errors in numbers or types of finished products have to be corrected at the most inflexible point in the supply chain. Any overstock errors occur at the point where all components have been committed to a finished item. If the item is perishable or subject to obsolescence, the reverse logistics operation required to recapture value is at its most expensive point. Overstocked product must be moved, disassembled, remanufactured, and restocked. Likewise, under-stock errors require the supply chain to react from the point of the longest leadtime. That is, product shortages must be made up from components. In addition, when finished items are placed at the point of customer consumption, forecast errors at that location create forward stock-rebalancing activities that add to overall supply chain cost. Thus, the advantage of quick response times comes at the cost of inevitable errors in providing the right product at the right place.
>From the producer's perspective, the CTO supply chain has numerous advantages. First, there is greater postponement than with a BTS supply chain-the producer need not commit to a final product until an order is received. The components and modules may require some precommitment but not the end-item configurations, because the order is entered at the preassembly point. Second, because the producer doesn't commit to the finished item, there is less aggregate inventory because there is less diversity of modules than of finished products. Third, the producer need only forecast and plan at the component level, as we will discuss in more detail later in the section on "CTO Rate-Based Planning."
In the case of build-to-order, the customer waits the entire time for the product to be completed. For some products this can take weeks, months, or even years. BTO products are often manufactured from order backlogs, which serve as the "shock absorber" for variations in manufacturing and demand rates. Because customer leadtime is exacerbated by the backlog wait time, BTO manufacturers sometimes move the orders of preferred customers forward in the order-release schedule. However, such order "shoe-horning" disrupts the capacity planning of the manufacturing facility.
The BTO manufacturer has less speculation risk compared to BTS and CTO manufacturers because fabrication is not committed prior to a firm order. Although this feature benefits producers and their suppliers, it provides few benefits to customers. The customer accepts a long leadtime as a painful necessity in order to benefit from a high degree of customization. In addition, customers served by BTO supply chains are required to forecast their requirements over the leadtime fulfillment interval. These forecasts are often inaccurate, requiring order and delivery adjustments prior to promised delivery date. For example, at one BTO plant we noted that finished goods were sitting in warehouses and railcar sidings. Upon inquiry, the plant manager stated that they really didn't have finished-goods inventory because they built to order. Yet the physical evidence was before us. The finished goods were present because customers delayed receipt of goods as the due date approached. That is, customers' real-time needs caused them to request a delay in shipment. The customer's forecast error became this plant's finished-goods inventory.
In general, the trade-offs between BTS, CTO, BTO, and ETO supply chains can be characterized along five critical dimensions, as shown in the "Supply Chain Design and Value Trade-off" box on this page.
*Toward the Ideal Supply Chain Design *
The phrase "maximize external variety with minimal internal variety" succinctly captures the principle that managers should follow when designing any supply chain.3 In other words, the ideal design is one in which a small number of components are used to configure a large variety of end products. This principle can be accomplished by structuring product offerings so that material and resource commitment is postponed for as long as possible. We will refer to this as the RAP (or keep in-process inventory as "raw as possible") principle, which is shown in Exhibit 2. The RAP principle is essentially realized by the CTO supply chain, which differentiates the product only at the final-assembly stage. The extent to which the RAP principle can apply to a supply chain's design is often determined by the configuration of the item being produced and the leadtime requirements of the customer. But even consumer packaged goods that use a BTS system can still benefit from the raw-as-possible principle. For example, a bottling facility can employ the RAP principle by using in-line labeling of two-liter bottles rather than purchasing prelabeled bottles from suppliers. The product configuration is typically captured by its bill of materials (BoM). Therefore the first step in any supply chain design or redesign effort involves configuring or reconfiguring the BoM to facilitate the RAP principle. Companies can achieve an RAP supply chain design simply by pulling unique parts that are aggregrated at different BoM levels, or different points in the assembly process, to the same level across different end products. For example, in the apparel industry, the identification of a specific brand name (say, with a decal) can be done further downstream simply by altering the BoM accordingly. This postpones the point at which these unique parts (decals) are assembled onto the end products, thus delaying differentiation according to the RAP principle. Closely related to the RAP principle is the principle of aggregation, or risk pooling. It is well known that aggregated demand has lower variation than demand for individual products. In some instances, the BoM can be optimized to exploit risk pooling. The idea is to pull unique materials to the same location (BoM level) across multiple stock-keeping units. By so doing, the producer can strategically locate safety stock upstream to pool the risk across individual product BoMs.
*Shifting from One Design to Another *
Supply chain designs should respond to customer values rather than to being inherently appropriate for a particular type of product. That is, the product's structure and supply chain design should respond to changing customer requirements instead of being taken as a given, frozen in time. Indeed, we have observed two interesting shifts in supply chain design that align with the RAP principle: a marked shift from BTS to CTO designs and a similar movement from BTO to CTO. Shifting from BTS to CTO. As customers seek products tailored more closely to their needs, producers are widening their product offerings by extending their current product lines and adding new ones. There is, however, a trade-off between the scope of products offered and the resources required to support that scope. This trade-off is more severe for BTS supply chains. That is, a BTS supply chain requires greater "asset intensity" to support downstream distribution, inventory, and retail facilities than CTO or BTO supply chains. This asset intensity, in turn, must be supported by larger profit margins. The CTO model can provide better margins for many products because it often requires less asset and expense commitment to support downstream supply chain elements.
A number of industries are showing this shift from a BTS to a CTO model. Already, the recording industry is letting consumers configure music CDs by choosing their own songs. In the North American automobile market, some companies are experimenting with CTO supply chains for consumer-configured automobiles. Many consumer-electronic s products can be assembled and distributed quickly if a CTO supply chain is deployed. As Dell begins to move cautiously into the consumer-electronic s markets, we expect to see the segment migrate more quickly toward a CTO model. Shifting from BTO to CTO. The shift is even more dramatic as more traditional BTO companies are exploiting the RAP principle and moving toward CTO solutions. We have observed trends in this direction for the last five years across a number of industries. Historically, many BTO companies have argued that fast-cycle supply chains do not apply to them because they are highly customized "job shops." However, that stance incorrectly views the supply chain from the perspective of the final product. Many BTO supply chains actually use many common components, ingredients, and modules that are sufficiently repetitive for them to consider moving toward a CTO solution.
One company we know well-a producer of electrical conduit-has made that move. This company specializes in customized solutions for laying electrical conduit within commercial buildings. Every order used to be treated as an engineer-to- order job; as such, it was processed from custom drawings. Yet the company uses the same types and sizes of conduit on every job. We proposed that the company design the conduit in sections of standard lengths and angles. A software program could then arrange the custom-conduit layout using standardized elements for more than 95 percent of the job. The standard sections could then be pre-staged as uncommitted parts available for configuration. By applying this CTO supply chain design, the company's order-to-delivery cycle time was reduced by more than 80 percent.
A typical objection to such an initiative is that it creates greater inventory risk. Although that claim is technically true, the consequences are not as severe as might be expected. Process inventory is certainly required, but it is often less than the sum of in-process jobs crawling through the manufacturing facility in a traditional job-shop environment. In addition, the amount of in-process parts is matched to the estimated build rate per time period. If the planning and execution time cycle is fast enough, then the amount of pre-committed parts can remain relatively small, thus reducing the inventory risk.
*Relationship Between Design, Planning, and Cycle Alignment *
The supply chain design is strongly influenced by the product structure. In turn, the product structure influences both rate-based planning4 and cycle alignment. Rate-based planning is the mechanism for aligning capacity with execution in a lean supply chain, and cycle alignment is the degree to which demand and fulfillment are matched in time. These mechanisms are familiar to most supply chain managers, but amid the everyday challenges, it is easy to underestimate their significance to supply chain design. To show that impact, we will discuss the two mechanisms in more detail in context of BTS and CTO supply chains.
*BTS Rate-Based Planning *
Typical BTS producers use planning bills of materials to forecast replenishment requirements for finished products and components.5 With planning focused on the end items, the BTS manufacturer uses the end-item demand rate to estimate the rate of component replenishment. The greater the scope of the end items, the more inventory is required to support immediate response. Put another way, the cost of speculative end-item placement puts limits on the scope of end items that can be supported economically. For example, a beverage bottler can provide various beverage flavors and container sizes but cannot economically support finer classifications of flavor or size. While the beverage producer can provide several sizes of soda, it cannot provide additional single-ounce increments from five to 16 ounces without exceeding the inventory and space levels it can reasonably and economically support. Thus, there are trade-offs between scope and economics for the BTS producer.6
The BTS supplier uses the planning bill of materials to design the supply chain capacity requirements. Components are speculatively placed to support upstream replenishment signals. For example in a soft drink supply chain, the periodic consumption of finished goods forms the basis of the bottling facility's production schedule, as illustrated in Exhibit 3. In this figure, two-liter bottles of a soft drink has a forecasted demand rate of 5,000 units per time period. This forecast is used to determine the demand rate and the associated demand variation for each end item. For this example, the three-sigma standard deviation of the forecasted demand sets the boundaries of expected demand between 4,000 and 6,000 units, or plus or minus 20 percent of the forecasted consumption rate. Thus, the forecast establishes the pacing rate as well as the expected variation to be accommodated by the bottling lines in a given time period.
The relationship between the finished product and its ingredients (components) is established in the bill of materials (ingredient card). For example, if the ingredient card requires one pound of concentrate for every 500 units of two-liter cola, then the concentrate demand rate is established at 10 pounds to support the forecasted rate of 5,000 units. The concentrate rate boundaries are set at plus or minus 20 percent, as with the finished product. The demand rates for the remaining ingredients are evaluated similarly.
In this way, product "streams" can be planned upstream from the point of order fulfillment (finished goods). As goods are sold, replenishment signals move upstream to replenish the goods sold. The supply chain is able to respond because the stream capacity is in place to satisfy actual demands. One can imagine that this process of linked pull signals would ideally move all the way back to the suppliers of the bottles, for example. If the beverage company could configure the supply chain so that demands are satisfied from a centralized warehouse, with fast transportation replenishment using "milk runs" and in-transit transfers of product, then the aggregate inventory could be reduced. Consolidation confers the benefits of reduced inventory by pooling the demand variation of the forward stocking locations into a single location. This same result can be obtained by employing advanced shipping notices to support cross-docking at the forward stocking location. In this design, the advantages of transportation economies are achieved without committing the forward stocking location to stock levels to replenish demand. The demand is replenished upstream, while the forward stocking location merely breaks, cross-docks, and stages consolidated shipments into store-level runs.
*BTS Cycle Alignment *
The BTS supplier must also align what we term the "wheels" of the supply chain-that is the production, delivery, and demand cycles shown in Exhibit 4 on page 56. The wheels capture the time within which the supply chain can respond across the complete range of the products. Although this concept is understood by most supply chain managers, it is often not given sufficient emphasis. We regularly discuss the need for inventory to buffer variation, but rarely do we discuss the need for inventory to support cycle imbalances, which is usually much more important. In the case of supply, the cycle is measured as the amount of time required to provide all the components to support the product range; for production, it's the amount of time the production system is able to produce all the products; for delivery, it's the amount of time between delivery dates for all the products; and for demand, it's the amount of time to sell the complete range. The BTS supply chain principle is to align these wheels to the demand wheel, as shown in Exhibit 4. Assume, for simplicity, that the BTS beverage company packages five different flavors of beverage in a single shipment. Now suppose that all five flavors are purchased across the retail network every day. However, moving one step back in the supply chain, if the product is delivered every three days, then the retailer must stock three day's supply of each flavor on the shelf to avoid stockouts between deliveries. Since the delivery wheel is not aligned to the demand wheel, inventory must result. The preceding discussion underscores the retailers' need for frequent deliveries to ensure that the delivery wheel more closely matches the demand wheel, thus reducing shelf space requirements for a particular SKU without compromising fill rates. The discussion also highlights the benefits of using third-party logistics providers (3PLs). 3PLs can provide aggregated transportation logistics across multiple suppliers to support increased delivery frequency and scale economies.
Now consider the bottling facility. Assume that the bottling facility is only able to produce the five flavors every seven days because of long changeover times and large batch sizes between flavors. Therefore, the bottling facility will need to hold multiple days of finished-goods inventory to support the delivery trucks. This view of the supply chain makes the value of improvements very obvious. If the beverage supplier could create a more flexible bottling facility so that every flavor is produced every day, and the delivery system could be made to deliver every flavor every day, then the whole supply chain would be fully aligned with the demand wheel. Such a configuration would minimize the amount of "cycle" inventory to the amount needed to support a single day rate plus safety stock to cover daily variation. Such a supply chain would flow to the rate of demand.
The speed with which forecasted demand is adjusted is a function of the sales and operations planning (S&OP) cycle. For example, a monthly planning cycle provides only monthly changes in the planned build and replenishment rates. Any change in forecasted demand must wait for new monthly signals from the S&OP process before the supply chain rate can be adjusted. Such long planning cycles usually generate large rate changes since demand changes are made over a monthly cycle. As the planning cycle improves to weekly or daily, the underlying planning rate can adjust more quickly to demand changes. As the planning cycle moves toward weekly and daily cycles, the supply chain can adjust more smoothly to demand changes. This adjustment must, of course, be coupled with increased flexibility within the supply chain to accommodate the revised rate broadcasts. Such flexibility is a function of engineering variables (constraints) inherent within the process. For example, for now the automobile industry can only economically adjust build rates biweekly, because of the cost of changing the assembly line.
*CTO Rate-Based Planning *
The planning structure for a CTO supply chain is designed such that there is less variety in components than in end-items. The supplier does its planning at the module level. For example, Dell does not plan using the demand-rate forecast for its finished products; there are too many possible variations for that to be worthwhile. Rather, Dell uses order management information to forecast module- and component-demand rates and variation. That way, its aggregate demand and variation can be determined at the component level by summing component averages and pooling variances across bills of materials. The upstream suppliers can use that rate information to provide immediate replenishment within the planned capacity stream. The order management system would track actual component and module consumption as orders are configured by customers. If the component demand exceeded the planned maximum capacity, the order management system would respond with an exception. The customer would be told that the component selection is not available-to- promise within normal delivery leadtimes, and an extended-promise date would be offered. The customer could then select the extended-promise date or select a different component that may not have exceeded planned capacity. Because the CTO producer builds items to customers' specifications, it cannot afford to lose leadtime by waiting for orders to aggregate into an economical shipment size for a particular region. The producer can gain logistical economies by combining custom orders into a single shipment. Such a strategy would reduce the amount of time any particular order waits for an economic shipment size.
*Continual Review Needed *
There is no one-size-fits- all supply chain design. Nor is there any guarantee of permanence for a supply chain design. Rather supply chain designs need to change as market sectors are buffeted by shifting customer needs, by unexpected moves from competitors, and by many other factors both internal and external. We have outlined four fundamental supply chain designs and explained the advantages and limitations of each. Each differs in terms of rate-basedplanning methodologies, order-point initiation, stocking strategies, degree of speculation risk, RAP potential, and push/pull signal placements. We have also pointed to important shifts from one design to another and provided additional viewpoints for examining supply chain performance with fresh eyes. Supply chain design cannot remain static. Just as the Internet has challenged preconceived notions of music distribution, so will technology, product, structure, and customer requirements change supply chain architecture.
It is our belief that supply chain managers expose their companies to unnecessary risk when they do not continually review the appropriateness of their supply chain designs. It is crucial that they evaluate competitive dynamics and artfully select the correct design for the correct set of opportunities.
Source :James M. Reeve and Mandyam M. Srinivasan
Posts
