The Customization-Responsiveness Squeeze

The machine tool business is an exemplar of beleaguered U.S. high-technology industries. In the early 1980s, most machine tool companies responded to the recession by slashing output. This left the door open for foreign machine tool builders to establish a strong presence in the U.S. market by bringing in cheaper, standardized products, most of which were built to stock offshore.

As the recession abated, most domestic firms had difficulty regaining significant market share in standard machine segments and concentrated on selling high-end and customized models. Customers had become used to the quick delivery available with the industry’s basic machines and, responding to their own need for faster deliveries, pressured machine tool builders for more rapid response. Most machine tool builders found themselves forced to produce a wide variety of machines in relatively small volumes, offer some degree of customization, and deliver with unprecedented rapidity.

Typically, build times are ten to fifteen weeks longer than desired delivery lead times. Therefore, a firm must have a machine well into the production process before it can be offered to customers with a competitive lead time. Because many of the expensive components cannot be used in other machines or even for alternative models of the same machine type, a firm is heavily committed to its forecast. Inevitably, the forecast is not completely accurate, and firms have had to develop tactics for coping with the consequences.

This problem is not limited to the machine tool industry; it is becoming common in many manufacturing and service firms. Developing the right software configuration for a system or delivering an automated assembly line are typical of the industrial activities that are being affected. In the past, customers were willing to wait for their specialized needs, but now customers are seeking out vendors that can develop, produce, and deliver near-customized products in record time. This trend has placed unique demands on organizations, and they must respond to the challenge or risk losing their market share in the important custom-product market.

We refer to this new trend as the “customization-responsiveness squeeze,” characterized by the need to deliver differentiated products in considerably less time than it takes to make them. Few companies can deal with this squeeze without modifying some key aspects of their marketing and production methods. The characteristics of this environment are as follows:

• the product is very expensive;
• volumes of products with any particular combination of features are small, and their demand is difficult to predict;
• the time to purchase components and manufacture the product (the build time or lead time) is much greater than the delivery time that customers expect;
• customers desire customized features of the product that must be established early in its build cycle; and
• customers are unwilling to bear the vendor’s cost or risk of carrying unsold units.

A common response to these conditions is to adopt the risky approach of continually initiating production of specific designs well in advance of receiving orders for them. We refer to this response as the build-to-forecast (BTF) approach. This is not the only available approach, however. There are a number of tactics that firms may employ either to avoid having to use the BTF approach or to reduce its risk if it cannot be avoided.

Figure 1 illustrates the range of possible production environments and responses. The production environment for manufacturing operations is largely determined by three dimensions: the delivery lead time demanded for the product (“responsiveness”), the desired amount of customization, and the stage of the production process at which the customization or differentiation can be effected. Two of these dimensions, delivery lead time and amount of customization, are customer driven; the third dimension, differentiation stage, is driven by the product’s design.

Typically, a firm will choose one of four basic approaches, according to the environment’s demands. If the amount of customization is low (the left half of Figure 1), the firm can usually employ a make-to-stock (MTS) approach and then use inventories of finished goods to provide short lead times. The inventory decouples production from the immediate demands of the market, reducing the importance of the stage at which product differentiation takes place.

High customization is where the problem begins. Production systems designed for high customization are usually radically different from those designed to support standardized, quick-response products. They would be highly inefficient if they tried to emulate MTS operations. But if customers are willing to wait for customized products, high-customization firms can use the make-to-order (MTO) approach (the lower right portion of Figure 1). With this approach, the firm commits to the production of specific products only when it receives orders for them. As long as customers are willing to wait, the stage at which product differentiation occurs may not be important. In fact, there may be little reason to have much product commonality, as each product can be produced independently.

However, the growing pressure for fast response and customization poses a new dilemma for firms. Neither the MTS nor the MTO approach is suitable. But if the product differentiation stage occurs late enough in the production process (the rear half of Figure 1), the firm can employ an assemble-to-order (ATO) approach. This approach allows the firm to make basic components to stock and then assemble them in various combinations as orders arrive.

If the product is differentiated early in its production process, the problem is more acute. This occurs frequently where product characteristics are largely established by fundamental components, such as frame sizes or power sources. Under such conditions, firms usually must maintain production systems that can provide high levels of customization, but they must modify these systems to respond to orders more quickly. Typically, these firms try to gain greater flexibility by enhancing their customizing abilities, reducing their build times in some way, or redesigning their products to allow differentiation to occur at later stages of the build process.

However, even these steps are not enough for firms in some industries, where competitive delivery lead times are far shorter than product build times and where stocking major components is not possible. There are, for example, industries where the initial manufacturing steps rely on materials that are themselves custom-produced and subject to long lead times (e.g., large specialized castings, specialty steels, or custom-made microchips). When all efforts to reduce build times still cannot meet customers’ demands for quick delivery, firms are now resorting to the build-to-forecast (BTF) approach.

How can firms better manage the customization-responsiveness squeeze? In this paper, we present a framework for analyzing the problem and choosing from a range of responses. As shown in Table 1, the framework consists of three steps: analyze customer expectations, assess the firm’s capabilities, and select and implement appropriate tactics. Throughout, we use examples from the machine tool industry.

Addressing the Customization-Responsiveness Squeeze

Before a firm can choose an appropriate mix of approaches, it must analyze its customers’ expectations and assess its own capabilities.

Customer Expectations

In the last decade, customer expectations for customization and responsiveness have changed. In industries where foreign-based competitors have concentrated on selling lower-priced, standardized products, many U.S. manufacturers have attempted to exploit their design and development expertise to broaden product lines and offer greater variety. By producing models at the technological leading edge, they could benefit from higher profit margins and greater differentiation. New technologies such as flexible manufacturing systems have been touted as solutions to the problem of producing customized units efficiently and quickly.1 As a result, customers in these industries have begun to expect increased levels of variety from all competitors.

At the same time, the success of just-in-time manufacturing methods and lean production systems has sensitized manufacturers to the value of reducing both build times and delivery lead times.2 Reducing production and purchasing lead times not only slashes manufacturing costs but also allows customer orders to be filled much more rapidly, which leverages the cost advantage. Most industries have thus come under pressure to meet new standards of delivery performance, regardless of their position in their supply chains.

These two trends are likely to continue. The Manufacturing Futures Surveys clearly indicate that customization and responsiveness are expected to keep shaping manufacturing competitiveness in the 1990s.3 Hence, a firm’s assessment of customer expectations in these areas is critical, and firms need to measure and monitor customer expectations on both dimensions.

The customization dimension is relatively easy to gauge because it measures the customer’s appetite for specific (as opposed to standard) products. Indicators of a market’s demand for customization are, however, frequently confused with a firm’s capability for customization. Just because a firm can deliver products in different colors doesn’t mean customers see much benefit in different colors. These are the critical questions to be answered:

• What are the customers’ perceptions of the benefits of customized products?
• Will the customer settle for something from a limited-range product line, with a standard product’s options and variants fulfilling the customer’s need?
• How much is the customer willing to pay for a particular amount of customization?
• How long will the customer be willing to wait for the customization?

The responsiveness dimension is somewhat more difficult to measure and interpret because it gauges the delivery time relative to the product’s build time. For example, a firm that must deliver a product in four weeks does not have a responsiveness problem if the product can be built in two weeks. Neither does it have a responsiveness problem if it builds the product in six weeks but its distributors agree to carry adequate inventories of the product to allow quick delivery. The real difficulty arises when customers demand a delivery time of four weeks, purchasing and manufacturing operations unique to the product take at least six weeks, and the risk of carrying unsold products is too high to inventory them as stock items. Thus it is important to assess the customer’s expectations about delivery time:

• What benefits do customers see in gaining quick delivery?
• What sort of price trade-off would be necessary to compensate for extended delivery times?

Price discounting may mollify customers who are looking for quicker delivery but, in many cases, not having the product available within a particular lead time may disqualify it from contention. Managers must have a good idea of the market’s sensitivity to this requirement — and recognize that the standard is probably changing.

At this point in the analysis, it is useful to consider the costs and risks of not satisfying customer demand in the highly customized markets. Perhaps the costs and the risks are relatively low. In that case, the firm may want to withdraw entirely from this market. But if we assume that the firm wants to continue serving customers with these highly specialized needs, then the next step is to consider the firm’s capabilities for addressing those needs.

Firm’s Capabilities

The firm’s capabilities determine which approaches to the customization-responsiveness squeeze it may realistically take. Managers should consider these areas:

• amount of excess capacity, both current and potential;
• cost and feasibility of carrying more inventory;
• availability and quality of product and process engineering skills;
• quality of planning and forecasting skills;
• value of current and potential technologies;
• cost and feasibility of altering the product’s differentiation stage to allow remaining work to be done within delivery times;
• feasibility of improving supply delivery; and
• acceptability of build-to-forecast approach.

Assessing these areas will give the firm a realistic assessment of its strengths and weaknesses and prepare it to make appropriate strategic decisions. Most firms do not assess many of these areas on a regular basis, making this step more difficult and time consuming. Some of the needed information, such as capacity estimates, are readily obtained. However, other aspects of firm capabilities, such as the feasibility of changing product differentiation stages, require specific, expert investigations. Further complicating the assessments here is the need to compare risks and benefits among the disparate areas that warrant investigation. Two things need to be stressed about conducting this step of the analysis. First, those making the assessments should view the firm’s systems as outsiders might, questioning current assumptions and methods. Second, teams drawn from different functional areas should conduct the investigations, not only to provide the needed expertise but also to add a variety of perspectives to the analysis.

A Range of Approaches

As suggested by the list above, the possible approaches can be organized into discrete areas: altering process design, altering product design, managing demand, managing supply, using slack resources, and building to forecast. However, these approaches are by no means mutually exclusive; combinations of approaches across different areas would offer synergistic benefits depending on the market’s expectations and the firm’s capabilities.

Process Design.

The obvious choice for most firms is to develop more flexible and responsive processes and systems. Flexible process design can be achieved through flexible technology, plant configurations based on the principles of cellular manufacturing, reductions in changeover times on conventional technology, and improvements in the production systems. However, these tactics offer only limited benefits.

The machine tool industry is itself a substantial user of some of its most sophisticated products, the flexible manufacturing cells and systems. With interconnected computer programmable machines, these systems can improve throughput times, allow wider part varieties, and provide high quality. These systems were once thought to provide manufacturers with the ability to generate new models faster and to produce small batches of parts with the efficiency of larger-scale operations, allowing a wider scope of outputs to be produced with equivalent economy. Unfortunately, most specialty machine production depends heavily on operations that are not as easily automated as the ones incorporated in flexible manufacturing systems. Assembly operations, metal casting, custom software development, and other vital production steps still defy programmable automation. Moreover, while the flexible systems in metal forming operations can be adapted to provide increased scope, attaining wider variety here can complicate downstream operations, especially in assembly. Efficiencies gained in the automated processes are thus often offset by the complications they create in other areas.

For instance, flexible manufacturing systems have proven their ability to provide a wide range of parts with speed and precision — parts that often lie idle for days awaiting components produced in the nonautomated operations. The use of flexible manufacturing technologies has not necessarily reduced delivery lead time; these systems are usually isolated “islands of automation” in the midst of processes that cannot be sped up unless product designs or manufacturing approaches are substantially altered. The lesson here is that manufacturing process technology alone cannot assure reduced production lead times.

A promising alternative is the reconfiguration of the plant according to cellular manufacturing principles. Setting up the manufacturing system in this way facilitates responsiveness for products with “controlled” customization. We have seen plants in a wide range of industries achieve remarkable improvements in delivery performance and build times by adopting cellular manufacturing.

Another effective mechanism is to use techniques to cut down production changeover times. Reducing setup time allowed one firm to increase its output level. The surplus thus created was used innovatively for responding to special demands from customers. In another case, setup time reduction allowed the firm to decrease batch sizes so that the inventory profile of work-in-process was much more balanced. Instead of having large inventories of only a few components, the firm now has small inventories of a large number of components. This allows it to be more responsive to specific customer requests while simultaneously reducing inventory investment.

Another tactic is to alter the information flows to improve their effectiveness. One firm added a formal liaison between marketing and manufacturing to facilitate communication about customer requests in one direction and operations restrictions in the other. Manufacturing is now evaluated primarily on volume, with the production system designed to minimize flow disruptions. A separate group in marketing handles major customization. The liaison manager specifically manages the conflict between the volume and customization orientation of the two groups. Another firm took advantage of the informality of its systems to make expediting more effective. It structured itself into a flat hierarchy (only two management levels beyond the plant floor) for dealing with responsiveness-related issues, thus minimizing delays on customer orders that required authorization for expediting.

Product Design.

One of the first priorities for firms facing the customization-responsiveness squeeze is to reevaluate their product designs. The best way to achieve product variety and speed is through modular product configurations. A modular approach can reduce the variety of components while offering a greater range of end products. Modularity can make flexible production systems much more effective, able to capitalize on economies of scope at key points in the manufacturing process. Although the time and cost of redesigning products may be considerable, the payoffs are great —not only in product delivery but in the firm’s ability to exploit such process technologies as programmable automation and just-in-time techniques.

One company redesigned its product line so that various models in the line share major subassemblies and components. This effort took a cross-functional team of product and process design engineers about fourteen months to complete. The modular components are usable on a wide array of models and therefore allow more substitution to match machines to customer orders. Problems still remain, though, if the forecast by basic variant types (established by the unit’s size and power) is inaccurate, because these aspects of the machines cannot be altered by simply switching add-on modules. Moreover, about 10 percent of all orders call for additional customization, further complicating the production process.

To benefit fully from product redesigns, the firm should simultaneously consider the production process capabilities also. This is an integral part of the “design for manufacturability” movement. The target here is to achieve the shortest possible production lead time and to push the stages providing product-differentiating features as late as possible in the production cycle. However, many firms may lack the engineering capabilities to accomplish this. Moreover, the biggest problem in redeveloping products for manufacturability is the great divide that often exists between the firm’s product and process design groups. That is why the firm just described set up an independent, cross-functional team specifically for the project. The team has since become a model for the rest of the firm as it attempts to gain similar efficiencies with other product lines.

Finally, automation of the design process (such as the use of computer-aided design systems) and standardized product design are important steps toward improving the firm’s ability to simultaneously compete on both customization and responsiveness. One firm achieved standardization by using product design engineers in sales. When a customer’s request was being processed, the design engineer was able to make substitution recommendations to the customer, thus reducing the inefficiencies generated by customization at an early stage. Automation of the design function allows early evaluation of the customer’s needs in terms of configuration and economic feasibility.

Demand Management.

The customization-responsiveness squeeze has put new pressures on firms to improve their forecasting abilities. Moving away from make-to-stock and make-to-order approaches means paying a higher price for forecast inaccuracy. Accuracy depends on obtaining good feedback from the field. This means that marketing departments may have to reskill their salesforces to keep their “ears to the ground,” anticipating customer orders well before receiving them. With the inevitable forecast inaccuracies, firms in this environment face the choice of either turning away customers or trying to alter the production schedule to suit their orders. The natural tendency is to choose the latter option. Unfortunately, this can easily lead to inefficient operations.

Two tactics help ensure that operations are buffered from forecast inaccuracies. The first of these is to switch the sales emphasis from reacting to customer requests to identifying possible customers for the available units. Regularly communicating the master schedule to the salesforce allows marketing to set its targets realistically and anticipate the units that are going to be available in the near future.

A second tactic is to establish “time fences,” where particular forms of changes to in-process units will not be accepted once they reach specific points in the production process. The firms we studied have difficulty accommodating production schedule changes because alternative components are rarely held in inventory. Altering specifications for partially completed units also creates confusion and delays throughout the manufacturing process, potentially jeopardizing the already tight delivery schedule. Formalized constraints invoked against schedule changes underscore manufacturing’s reluctance to accept them. Such measures are a compromise that protects manufacturing but still allows some flexibility to alter units in production.

At one company, high demand and an expanded product line led to an increased level of back orders, with growing numbers of mismatches between the units in production and customer orders. In response, marketing executives tried to get manufacturing to alter unsold in-process units to match unfilled customer orders. Manufacturing resisted this effort because this policy would require it to hold optional components in inventory, each of which might cost several thousand dollars. As a result, the company established a liaison staff to negotiate marketing’s requests for in-process unit modifications. Strict guidelines were drawn up delineating the points in the production process after which no modifications could be made.

Supply Management.

Three supply management tactics can help the operations function respond to customized requests quickly: (1) internally developing the capability to fabricate frequently ordered, long lead-time components; (2) altering purchasing contracts to provide for quick delivery; and (3) developing production management systems geared to external customer demand (referred to loosely as “ship sets”) instead of internal batch sets

To develop internal fabrication ability for critical parts, first identify the critical parts and their impact on delivery time. It is likely that relatively few items are of key importance. Then establish an internal group to produce these components quickly, that is, faster than the company can source them. It is important that the group’s critical performance measures be based on responsiveness and variety, rather than on cost and output volume.

Another tactic is to renegotiate contracts with suppliers to improve parts delivery. One particularly devastating source of uncertainty in the machine tool industry is castings, which take as much as six months to procure.4 In one firm, quick delivery was assured by vertical integration. In another, the firm negotiated to carry the supplier (that is, purchase from them at a fixed level) during lean demand periods in exchange for quick delivery. Another firm formally negotiated with a vendor to reduce the delivery lead time by issuing a long-term “blanket” order for an expensive part.

Another tactic that improves responsiveness is “ship set” production, that is, the production of small batches of parts in numbers that exactly match those needed for specific products. In doing so, components groups needed to fulfill a customer order are sped through the production process without having the longer processing times of larger batches. This tactic is much less disruptive when small batch sizes are practical, which occurs when setup times are reduced.

Slack Resources.

Excess inventory and excess capacity can directly mitigate the responsiveness dilemma in customized markets. However, these are expensive choices, and charging the customer for them can mean losing ground on price competitiveness.

Maintaining an inventory of finished goods or commonly used, long lead-time components may be affordable in markets with “controlled” customization. One firm’s policy is to carry a specialized item in inventory if its demand is greater than four units per year. A finished goods inventory of currently popular units is an easily implemented solution, but it risks high costs and losses through obsolescence.

Other resources may also be available to mitigate the problem. One firm in our field studies chose to use its surplus production resources during lean periods for expediting specific orders through the system. Another used its surplus engineering resources for major product redesigns; the team reduced product build times by nearly 35 percent and costs by nearly 40 percent. In each case, slack resources were committed either temporarily or on a continuing basis to absorb some of the impact of the dilemma. However, in each case, committing available slack resources to deal with the customization-responsiveness squeeze was not enough. Other measures also had to be taken to manage the problem more effectively.

Build to Forecast.

We found in our field studies that most companies could mitigate the customization-responsiveness squeeze by using some of the tactics described above, but they still needed to build some units to forecast and then to try to match these units as customer orders trickled in. The build-to-forecast approach focuses on the units that are close enough to completion that they are “available to promise” to customers at competitive lead times. If each period’s forecast is perfect, then all customers can be offered units within acceptable delivery times that match their requirements, leaving the manufacturer with no leftover units. However, forecast inaccuracies are common, especially with forecasts of small volumes made perhaps six to twelve months in advance. Customers then have to wait too long for the desired units or the company is stuck with units that do not match customer orders.

One solution is to freeze any additional work on unmatched units at some point in the build cycle, thus creating an island of work-in-process inventory. In most situations, however, unmatched units must continue along the production line: there may be no space to set aside extra units, and there may exist only enough material for the forecasted units, so that pulling an unmatched unit off the line results in a “hole” in the production flow that idles workers and machines at each station in sequence, possibly for days at a time.

Another solution, as we have noted above, is to hold unmatched, completed units as finished goods, allowing quick delivery if orders eventually do arrive. However, this is risky, particularly in high-technology firms that maintain competitiveness through rapid, continual product improvement; completed units may quickly become obsolete. Some firms “retrofit” completed units to match orders by disassembling and altering their configurations, a very expensive approach.

The build-to-forecast approach offers a feasible, although risky, solution.5 It should be considered only after all other tactics have been tried, and then firms should adapt their sales and manufacturing systems to reduce the inherent risks.

Choosing Appropriate Tactics

Every company has a unique combination of capabilities and market needs. This combination determines the firm’s choice of tactics. Let’s consider how three machine tool companies that are market leaders in their segments are responding to the customization-responsiveness squeeze.

Company A is a medium-sized, privately owned firm. One type of machine tool accounts for 70 percent of its business. Because of its narrow focus, Company A has been able to rely on a small finished-goods inventory of the main line’s standard models to buffer the effects of forecast inaccuracies. But this inventory does not always cover demand. Rather than expand this inventory, the company invokes other practices when neither in-process nor finished units will be available in time. If the order is for a low-end model, the customer may be induced to buy an available larger or higher-grade model. If not, then the company may offer the customer one of these units with the unwanted features removed or disabled. If the order is for a high-end model, a showroom unit of exceptional capability may be delivered on loan until the specified unit can be prepared. Only in very unusual circumstances will the production process be disrupted to accommodate a specific customer order. On such occasions, work on the special units is expedited to meet the delivery deadline. The company’s production process has considerable slack; expediting a unit through the latter production stages is feasible. In addition, the firm has few layers of management, and the manufacturing managers can easily approach the firm’s owner to get authority for the required expediting. To date, the firm has not considered altering its product designs to accommodate the lead-time pressures it is experiencing, relying instead on incremental product and production process improvements to keep the problem under control.

Company A can afford to take a fairly casual approach to this problem because of some important factors. First, its limited product line increases the likelihood that units and customer orders can be matched up by using either the small finished-goods inventory or available units in the final weeks of production. Second, the problem is mitigated by the company’s ability to downgrade some of its completed units to accommodate orders for lower-grade models, although this is a potentially expensive alternative. This product commonality also allows the company to expedite delivery lead times with relatively few organizational changes. Even so, Company A is concentrating on improving its production process with cellular manufacturing techniques to allow faster throughput. Unfortunately, it will still have to commit to specific model production on a forecast basis, since the total throughput time is expected to remain longer than competitive delivery lead times.

In contrast to Company A, Company B reflects the experience of some of the industry’s foreign-based competitors. It gained substantial market share through the introduction of its high-volume, standard models. It competes in several machine tool industry segments, producing a wide variety of machine types in a centralized facility.

Company B strives for level employment in manufacturing and has consequently developed mechanisms to buffer the production process from the industry’s demand swings. It has established a system whereby the manufacturing department builds machines according to a six-month forecast and then sells the machines to the marketing department. The manufacturing department is judged on its performance relative to planned output. In slow periods, marketing must absorb unsold units, holding them as finished goods or pushing them on to their distributors. In the past, this make-to-stock approach worked well with the limited range of standard units that were produced in high volumes. Moreover, Company B has sufficient financial resources to carry significant inventories of these machines.

But pressure to provide a faster response to requests for its nonstandard models has pushed Company B into broadening its range of domestically produced models. Its wider product variety means that B cannot rely on the simple steps taken by A. Company B is the one described earlier that established a liaison to negotiate marketing’s requests for in-process unit modifications and that pushed customization into a marketing-operated facility. The liaison system, the post-manufacturing customization, and the attempt to build a finished goods buffer are all directed at minimizing disruptions to established production plans. Company B’s management systems still resist the destabilizing influences of customization. As a result, even with the various mechanisms to provide for unforecasted units, the salesforce has learned not to expect many changes to in-process units and instead “sells to the forecast.” Sales efforts have concentrated on selling what is available or, in effect, locating customers who have specification and timing requirements that match the production schedule.

Company C is one of the industry’s oldest and largest firms, with considerable engineering expertise in both product and manufacturing process design. When one of its main product divisions began to experience considerable delivery lead-time pressure, its managers took an entirely different approach for improving its responsiveness — through redesign of both the products (to achieve greater component commonality) and the production process (to reduce production lead times).

Despite adopting a flexible manufacturing system and cellular manufacturing techniques to speed throughput, production times still exceed typical delivery lead times by about twenty weeks. However, like Company A, this division has one main product line, which makes for a simpler problem than the one facing Company B. This is the company described above that redesigned its product line on a modular basis. The modular design allows greater flexibility for matching in-process machines to customer orders, but the practice of changing unit specifications has injected some instability into the production schedule. On a few occasions, unmatched machines nearing completion have undergone major “teardowns” to alter them to suit unfilled orders for similar models. The division now has a policy of not quoting a delivery time of less than eight weeks in an effort to control the number of last-minute changes. Despite these measures to ease the order-matching problem, the division’s salesforce tends to sell to the forecast, concentrating to some extent on selling available units that are eight to sixteen weeks from completion.

These three companies demonstrate workable, if imperfect, methods that have allowed the firms to maintain leading market share positions. They also show that the severity of the problem tends to vary with the extent of customization offered by the firm (see Figure 2). Company A has relatively low levels of customization and can use relatively simple mechanisms and informal systems to deal with the problem. Company B builds inventories of standard products but has to provide systematic organizational mechanisms for its medium-level customization. Company C, with the highest level of customization, has adopted formal systems as well as new product and process designs.

In addition, the cases illustrate how unique resource capabilities affect the choice of tactics. Company C, with its engineering prowess, chose to redesign the product. The firm with a financial cushion, B, chose to build units to completion more often and had the highest finished goods inventory level. Company A, with some slack in production capacity, chose to expedite customer shipments more often. The adopted measures all involve modifying the management systems that trigger and control production.

Conclusion

The customization-responsiveness squeeze is likely to accelerate and to affect more industries. Often it is not easy to recognize the problem until a number of system changes have been made to deal with growing customer dissatisfaction. Managers should be aware of their business’s susceptibility to the dilemma and be prepared to take proactive steps to deal with it.

Each of our case companies made a deliberate choice to base the production of a wide variety of specialized units on forecasts. The alternatives were to limit their product ranges or offer nonstandard units on a make-to-order basis, which would have resulted in long delivery lead times. Each firm’s management felt that the build-to-forecast approach offered less risk than the certain loss of market share that the alternatives entailed. This will not always be the case. Managers who recognize that their business is being affected by the squeeze should weigh the potential effects of all alternatives before adopting any form of build-to-forecast approach.

Many of the industries adopting a build-to-forecast planning approach have been hard hit simply because of increased global competition. Customers who have previously had to wait for specialized manufactured products can now find alternative sources. With wider choices, customers continue to compare competing products on the basis of quality, price, delivery, features, and other dimensions. To compete, producers must meet minimum standards on each critical dimension. Few firms have been immune to the effects of increased standards for both delivery and variety within their industries. Even if a firm has products that are well differentiated and strong on one particular dimension, it must still respond to changes in customer expectations on other dimensions. A small division at Company C is a good example. Some of its models, which are all built for one industry, sell at the rate of only a couple units per year. Despite being one of the few producers of its type of specialized machines, this division had to conform to customer expectations that its machines should be available within lead times approaching those of standard machines, about half the build time. Otherwise, its customers might be willing to trade off the performance advantages of the specialized machines against the alternative of selecting less specialized machines that might be available several months sooner.

Regardless of the chosen response, there is one theme that runs through all the successful efforts we have seen — functional integration. Cross-functional teams, functional participation in strategy development and implementation, and organization structures that provide the latitude and flexibility to deal with multidisciplinary tasks will, we predict, be of increasing importance. These tactics will require changes throughout the firm.

Virtually any effective mechanism requires coordination across departmental boundaries — such as marketing and manufacturing, or product design and process engineering — that traditionally have not functioned well together in many firms. The approaches that we observed as having the most promise relied on cross-functional teams to plan and implement the changes. In addition, the coordinated efforts required long-range planning that reflected strategic input from marketing, manufacturing, and purchasing. No effective solution is likely to come from the isolated efforts of one department. And to achieve organizational change, managers must realize that the old methods and structures were rooted in a different competitive environment; new structures need to have the flexibility to roll with the punches as the firm adapts to competitive changes.