Discussions on vertical integration—about whether products, parts, or services should be produced in-house or outsourced—can often become heated exchanges between managers holding oppos-ing business beliefs. Is it a question of optimizing the cost structure, or destroying jobs? Of making fixed costs variable, or losing knowhow? Of gaining flexibility, or plunging into dependence?

Not surprisingly, the "make or buy" debate almost invariably runs into contentious areas. Some see in it an opportunity to thrust forward into a spectacular hi-tech investment. Others see it as representing a misguided sense of loyalty to the business’s in-house team. Criteria for an objective decision are neither easy to find nor easy to apply, since every case is different.

However, the good news is that make or buy considerations can be expressed in objective terms. What is not such good news is that these considerations have to go much farther and deeper than is normal or convenient. To be useful, changes in a company’s level of integration should make its entire value-added chain more effective—not just consist of spinning off units or adding on operations or processes in a piecemeal fashion. And when we talk of vertical integration, we include all elements of the chain, not just production.

High or low?

No amount of observation of prevailing practice in different industries will reveal whether high or low vertical integration creates competitive advantage. Neither will looking at whether successful companies tend to be less or more vertically integrated. Our research showed that in Germany, successful component manufacturers consistently do more in-house production than less successful ones, while in machinery manufacturing, some of the successful companies had high vertical integration and some low. Western auto makers are traditionally highly integrated, though now tending to reduce integration; conversely, in plant construction, the most successful companies have very low vertical integration.

Clearly, every company must find its own optimum level of integration. Vertical integration should be used as a means to reinforce existing effectiveness in technology or operations. Thus, if a company is technologically or operationally superior to its competitors and suppliers, a high level of vertical integration will give it a competitive advantage. If it is weaker, on the other hand, the same level of integration will be a disadvantage.

Outsourcing to compensate for operational or technological weaknesses only works if the elements that are farmed out have nothing to do with what differentiates a company in the market. Companies that outsource production simply because they are not cost-competitive must be prepared to be overtaken by more efficient competitors, or even by their suppliers entering the market in direct competition.

Vertical integration should be used as a means to reinforce existing effectiveness in technology or operations

Component manufacturers rarely achieve strategic differentiation with a superior product concept; their strengths are more likely to lie in operational excellence, larger economies of scale, a superior process, and first-rate logistics. As a result, the best among them tend to have higher levels of integration than do the less successful.

Machinery companies, on the other hand, are able to achieve adequate differentiation with a superior system architecture or design, and have no real need for particular operational strengths. To some extent, they can compensate for operational weaknesses by outsourcing. What distinguishes excellent machinery manufacturers is that they are a better judge of their competitive position than poorer performers. They aim for high vertical integration when they are strong in operations, but outsource other value-added stages when their superiority lies more in the machine concept or design.

It is open to question how long this latter strategy will remain sustainable. Today many machines become commodities as their life cycles advance, making differentiation possible only through low costs derived from operational efficiency. By this point, at the latest, a manufacturer must be in a position to exploit economies of scale and raise its competitive profile by means of superior operational management.

While reducing vertical integration is not a panacea for low operational efficiency, no company is operationally superior in every field

In some cases, such as the automotive industry, optimizing vertical integration actually means reducing it; only in this way can companies make the necessary 30 to 40 percent cost reductions. This can raise difficult social and political concerns. But ignoring the issue could prove fatal, given the strength of international competition. Creative solutions are called for. They might include such moves as spinning off component manufacturing facilities to create independent companies with their own development departments as profit centers, or contracting out entire manufacturing stages to suppliers.

While reducing vertical integration is not a panacea for low operational efficiency, no company is operationally superior in every field. Even outstanding operational management cannot compensate for the negative effects of lost economies of scale or high personnel costs.

Successful companies have found a clear line of their own in the muddled make or buy debate:

• In making their decision on vertical integration, they consider both technological differentiation and cost optimization.
• Once they have decided which elements should be outsourced, they ensure good cooperation with their suppliers through clearly defined system boundaries and such trust-building signals as a reduction in the number of suppliers.
• They instill throughout their organization the new procedures and modes of behavior that make cooperation and change happen and that determine, in the long run, whether day-to-day operations will succeed in realizing the benefits of optimization.

MAKING THE DECISION

Manufacturers in the United States have gone through phases of thinking that they could differentiate themselves simply with system concepts and packaging. Having consequently transferred large elements of their value-added chain to low-cost countries, they now realize that they have lost whole parts of their industrial base.1 Countries they once viewed as no more than cheap supply markets have now overtaken them in the marketplace.

This state of affairs provides clear macroeconomic confirmation of the power of vertical integration in reinforcing technological and operational effectiveness. Accordingly, the current revival of some US industries is based on a return to productivity in manufacturing, coupled with an increase in in-house value-added. In many cases, the Americans are now outperforming their German competitors, many of which are yet to experience the full force of Japanese competition.

Successful companies avoid losing technological and operational effectiveness by careful handling of two key issues. First, on the strategic side, they determine which activities have to be performed in-house because they are essential to technological differentiation (Exhibit 1). They then proceed to assess the cost advantages and disadvantages of producing in-house compared with using efficient suppliers. In doing so, they take an integrated view of costs: factor costs, scale advantages, productivity, and also—most importantly—the potential savings in design-to-cost.

Technological differentiation

The best safeguard of competitiveness is technological differentiation that offers clear value advantages to the customer. Companies must determine both whether there are new opportunities for technological differentiation and whether traditional sources of differentiation are still valid. Caution is required: for instance, a technological advantage might be sustainable, or short-lived. Two key questions can help determine whether in-house development and production is necessary for technological differentiation:

1. Does the sub-system or component have a decisive influence on the overall architecture and performance of the product? If so, it is a matter of survival for the company to have the knowhow and structures to develop and produce it competitively. In-house manufacture or development is vital for such core components of a product. Other important sub-systems do not necessarily have to be produced in-house, although the relevant development expertise must be retained. A manufacturer must be able to have a say in an outside supplier’s design, but without destroying the advantages of outsourcing through over-specification.

2. Are there competitive advantages to be won from a technological innovation for the sub-system or component? In-house development and manufacture should be considered only if its cost and technology advantages can be defended over time. Typically, this means setting up a separate profit center that is free—apart from restrictions on direct competition—to operate in the market and commercialize its products. Such an approach will work where the funds and management resources to build a strong parallel business are available—and where sufficient economies of scale can still be guaranteed if competitors establish collaborations with third parties. Often, though, close cooperation with a good supplier turns out better than going it alone.

Close cooperation with a good supplier often turns out better than going it alone

Companies must take care not to let the aim of technological differentiation absorb all their attention. Its advantages should not carry such a high price tag that they actually reduce value to the customer. When numerical controls were first introduced, for example, many machinery manufacturers tried to produce their own control electronics. As basic development costs for controls increased, and with them the importance of economies of scale in the hardware, these companies found it impossible to manufacture the controls competitively. Instead of a technology lead, in-house development created a severe competitive disadvantage in comparison with suppliers, which were able simply to add their own modifications to standard hardware and software.

Optimization of costs

Once the issue of technological differentiation has been resolved, the question of costs needs to be tackled. Frequently this turns out to be no more than comparing quoted prices against in-house marginal cost.

While such an approach may seem to make sense, it is not the best way to go about optimizing the complete value chain. External quotations may in fact be loss leaders from suppliers intending to raise their prices once the customer has reduced its level of integration. Or quoted prices may be inflated because suppliers have based their costing exclusively on existing drawings, without making any allowance for optimization potential. At the same time, they often try to write down any special fixed assets over a very short period, being reluctant to rely long term on non-binding agreements on order volumes.

Neither is in-house marginal cost an appropriate basis for comparison. When a decision on vertical integration is being taken on strategic—and thus long-term—grounds, it makes a lot more sense to base it on the full cost of production. Even this falls short: the cost of development and production engineering should also be taken into account.

The issue for management, then, is to understand the real cost differences between manufacturer and suppliers. There are four areas to consider:

1. Factor costs. Suppliers’ factor cost advantages, particularly in personnel, can be so great that other advantages inherent in in-house production dwindle by comparison. In the automotive industry, for example, suppliers’ personnel costs are typically 30 to 40 percent lower than those of OEMs.

Successful examples of global sourcing are few and far between

Superficial analysis based on annual labor cost per employee often shows supplier personnel cost advantages of a mere 10 to 20 percent. A more useful benchmark, however, is total annual labor costs—including such ancillary costs as unemployment insurance—divided by number of hours worked, yielding net labor cost per hour. Since in Germany suppliers’ employees often work 1,500 to 1,600 hours per annum, while OEMs’ employees chalk up only 1,100 to 1,300, this calculation gives suppliers further cost advantages of 20 to 30 percent. Add this to the lower annual labor cost per employee and the personnel cost advantage per hour reaches 30 to 40 percent. For countries with low labor costs, it may be even higher.

In view of these cost advantages, many manufacturers might consider the make or buy issue already settled. But that would be wrong. A too hasty subcontracting of parts and systems to suppliers can quickly lead to substantial cost disadvantages—the result of inadequate supplier skills in development, delivery capabilities, or quality.

This may explain why successful examples of global sourcing are few and far between. On the one hand, to exploit suppliers’ factor cost advantages, a manufacturer must define the technical interface—the point at which a supplier’s contribution fits in to the manufacturer’s own product—so clearly that there is little need for further coordination. On the other, it must play an active role in closing suppliers’ capability gaps.

Global sourcing currently holds opportunities for US manufacturers in Mexico and for European manufacturers in some former Eastern bloc countries that are now heading toward a market economy, such as the Czech Republic, Hungary, or Poland. But this approach will succeed only if manufacturers provide funds and management resources to bring development and production knowhow up to Western standards, or if they can induce their traditional Western suppliers to provide such "development aid" through joint ventures or long-term purchasing guarantees.

The attractiveness of low labor costs should not blind companies to the possible risks

The attractiveness of low labor costs should not blind companies to the possible risks. Most development departments in low-cost countries are far from world-standard, despite a sound technology base. Production plants are often obsolete and need replacing, ideally with greenfield facilities. And management should also bear in mind transport problems (and their impact on production schedules), political risks, and, in years to come, rising wages.

2. Economies of scale. Suppliers’ scale advantages are often as great as their factor cost advantages. Clearly, a supplier will usually have a much higher unit output. But what is often not realized is that economies of scale may also exist in material costs, especially if a manufacturer is paying more for all the various components together than it would for a complete outsourced system. A system supplier will probably get much better discounts on its bulk purchase of individual components.

3. Productivity. Manufacturers often lag well behind suppliers in productivity. In order to measure total productivity in a facility where up to half the staff work in non-manufacturing functions, a benchmark needs to be developed that includes indirect labor. Among its components will be the total number of hours (direct and indirect) per product and—as an indicator of how "lean" a manufacturing organization is—the number of productive hours divided by the total number of employees.

Here, too, suppliers frequently enjoy a 20 to 30 percent advantage. The reason is simple: with their organizational transparency, owner-based culture, and entrepreneurialism, they tend to work with much more efficient structures and be far more productive than large manufacturers.

However, companies should not jump to conclusions on the basis of pure productivity disadvantages that cannot be attributed to factor cost or scale disadvantages. Ensuring adequate productivity is a management task that must be addressed by means other than cutting out more and more value-added stages. But companies concerned about the cost of employees should be warned against the flight into over-automation. It has been known to precipitate ruinous investment decisions that achieve, at best, only marginal productivity improvements.

4. Design-to-cost potential. Factor costs, scale effects, and productivity differentials should not be the only criteria for decisions on outsourcing. It is just as important to estimate the potential for further cost optimization. One fruitful source is logistics. A manufacturer needs low materials inventories; not only that, but total warehousing, transportation, and setup costs also have to be minimized for both manufacturer and supplier.

Perhaps the greatest of all scope for improvement lies in design-to-cost

Perhaps the greatest of all scope for making improvements lies in design-to-cost. As a rule, this potential is available only to the organization with the greatest experience in systems optimization and the best knowledge of the manufacturing processes involved. That may be the manufacturer, but it is often the supplier, especially when important sub-systems are involved. The amount of costly platinum needed for an automotive catalytic converter, for example, was considerably reduced in one case when a supplier’s experience in flow optimization was applied at an early stage of development.

Design-to-cost potential is particularly high where a supplier can tailor a design better to fit its own manufacturing conditions. One machinery manufacturer, for example, had always insisted on the use of a forging in one of its sub-systems. However, its supplier was able to satisfy the same requirements with a sheet metal part that it could produce much more cheaply with its existing plant—thus cutting costs for both parties.

Exchanges of experience between manufacturer and supplier often bring to light unexpected opportunities for making improvements. One German manufacturer sent a standard part to Japan for redesign and a comparative quotation, and initially had its expectations confirmed: the Japanese supplier could deliver around 30 percent cheaper. Its surprise at the second attempt, with a much more complex part, was all the greater. The Japanese supplier was some 18 percent more expensive than the company itself, despite the redesign.

Yet it would still have been wrong not to outsource—as a more detailed analysis of the costs revealed. For the Japanese supplier, integrating the complex part into its manufacturing process caused major problems, as was made evident by its superior cost accounting system, which showed

the full cost of complexity. At its German customer, by contrast, the simple parts had been subsidizing the complex ones. Once the costs had been properly allocated, the Japanese alternative proved to be more cost-effective in the long term than in-house development.

A closer look at the Japanese supplier’s redesign brought further rewards. On its own, the complex part offered hardly any optimization opportunities, being interconnected with too many other components. But optimizing the total system—redesign and fresh development of all the components by the Japanese supplier—revealed the full potential for cost reduction.

In both these examples, design-to-cost potential could be exploited only via increased subcontracting. But outsourcing production can also have its detrimental effects. One heavy engineering company decided ten years or so ago to reduce its vertical integration and stop making parts altogether. Since demand fluctuated widely, the manufacturer assumed that converting fixed costs to variable costs would make it more competitive.

While this plan looked good at first, the drastic cuts almost destroyed the company. Because the cuts were made in machine design, the developers lost their understanding of manufacturing processes, and so their product/ process optimization deteriorated. At the same time, the company’s suppliers were not prepared to optimize their own manufacturing processes to fit the products, since they never knew when the next order was coming or whether their contracts would be renewed.

If a part neither serves technological differentiation nor yields cost advantages, its manufacturer should outsource

Amid such uncertainty, it was impossible to optimize procedures and processes. Surprisingly, the company did not fall too far behind technologically, and it remained perfectly competitive in the decisive system technologies. On the cost side, however, things looked less rosy. The nature of the interface between development and production made discussion and coordination difficult, with the result that design-to-cost approaches were forgone and extremely high opportunity costs were incurred. The company’s overall cost position became uncompetitive.

As a rationale for vertical integration, technological differentiation is valid only in conjunction with a detailed cost analysis. If manufacturing is to be kept in-house because of differentiation, it is essential to make its total cost transparent and suitably competitive. If a particular part neither serves technological differentiation nor yields cost advantages through in-house development and production, its manufacturer should switch to a strict policy of outsourcing. Here, getting the right form of collaboration with the supplier is at least as important as getting the original decision right.

Cooperation with suppliers

Unilateral action will not take a manufacturer very far toward exploiting the potential for improvement it has identified. Realizing both design-to-cost potential and cost advantages depends on achieving the best form of cooperation with the right suppliers. Continuing to operate in the same old ways can jeopardize—or even reverse—the advantages.

Successful companies also simplify their relationships with suppliers by creating trust

The guiding principle is to keep things simple. The interface with the supplier should be designed such that a clear demarcation of the system boundary simplifies the operations of both companies and, at the same time, opens up possibilities for technical improvement. Successful companies also simplify their relationships with suppliers by creating trust. When both parties know they have a fair deal from the outset, their energies can be applied productively to shared improvements rather than wastefully to a host of checks and controls.

System boundary

The right system definition is the key factor in the entire outsourcing process. Mistakes made here cannot be compensated for even if all the other conditions are fulfilled. The criterion for the right definition is simple: What system will enable the manufacturer to achieve both technological differentiation and optimum exploitation of factor, scale, productivity, and design-to-cost potential? In practice, two questions predominate: What interface should be chosen? and How large should the subcontracted system be?

Interface. There are three types of interface from which to select: families of parts, functional modules/systems, and assembly modules/systems. Whichever the case, a manufacturer will buy more from suppliers than it has in the past, and its focus on one or a very few suppliers will allow a new form of collaboration. Which interface to choose depends on the potential available:

• If it decides on a family of parts, a manufacturer will stop buying, for example, its various actuators or hydraulic units from several sources, and purchase them all from one supplier as a family of parts. Since the supplier will gain an overview of the manufacturer’s needs, it will become a more competent partner. It can help the manufacturer open up many more cost-saving opportunities by standardizing its various control or drive units. Both manufacturer and supplier can then achieve considerable reductions in specification and control costs.

  This approach will not normally reduce factor cost. Savings will flow mainly from economies of scale resulting from standardization, and to some extent from design-to-cost potential. For the manufacturer, supplier development becomes simpler, as it has to develop only one supplier instead of several.

• At first sight, the most sensible arrangement is to subcontract complete functional modules/systems. Here, the interface between supplier and manufacturer is defined by the specification of functional requirements and installation conditions. Within this definition, there is freedom for the supplier to shift various functions from one component to another, for example. Managing the interface is problematic only if the functional module is "distributed" over the entire machine. A risk then arises that over-complex coordination will be needed, that design-to-cost potential will not be exploited, and that functional problems may occur.

  Subcontracting complete functional modules/systems makes sense only when the supplier knows the system architecture better than the manufacturer does, possibly because it supplies a large number of manufacturers.

• The biggest savings are normally captured when complete assembly modules/systems—which should ideally also be functional modules—are outsourced. The merit of this choice is that the manufacturer only has to handle the module once, for final assembly. Much of the assembly value-added is shifted to the supplier, resulting in improvements in personnel cost, economies of scale, and design-to-cost. It is also easier for manufacturing and development to draw up specifications for assembly modules/systems than for families of parts and functional modules.

  Almost every machinery manufacturer has experienced the benefits of buying complete assembly and functional modules. Many have cut costs by switching from the purchase of individual gear wheels to complete standard gears, for instance, and some suppliers of active spindle-bearing units now offer products far superior to most in-house alternatives.

The biggest savings are normally captured when complete functional assembly modules are outsourced

The search for an appropriate module to outsource should start with the largest possible units of sub-assemblies, and proceed to progressively lower levels of aggregation if no opportunities are found. In extreme cases, buying entire product groups from external sources has actually proved wisest—provided both that the technology is extremely mature and no further technological discontinuities are on the horizon, and that the manufacturer is prevented by low volume from achieving differentiation with its process technology. If subcontracting of entire products or sub-modules works successfully, there is more chance of making largely residual fixed costs flexible after all.

Scope of supply. Once the interface has been defined, the issue becomes one of scope of supply. If the scope is too narrow, a supplier will have little room for optimization; if it is too broad, there is a danger that no supplier will have the resources to exploit the potential.

A generous scope makes sense when the system is largely separable—in other words, when the number of interfaces with other systems and sub-systems is limited, and when the connections at these interfaces can be precisely stipulated.

It is important to have a clear understanding of a supplier’s skills when deciding how much scope to give. When the scope is determined by logistical requirements, it can put too much strain on the supplier. For example, if a supplier that produces only electronic components is required to deliver both electronic and mechanical parts, it will quickly come to regard the latter as an extra for whose costs it does not feel responsible—and those costs will soon start to rise.

A supplier can only achieve better system costs than a manufacturer if it develops and manufactures a large proportion of the components

A supplier can only achieve better system costs than a manufacturer if it develops and manufactures a large proportion of the components. By doing so, it can optimize entire groups of parts or aggregate individual functions, thus making quantum leaps in cost or performance. It is conceivable—particularly with mass production—that suppliers with different strengths will form a project-specific or multi-project cooperation and offer good package deals. Manufacturers need to be careful here, though, since purchasing a system from two cooperating companies makes much greater demands on all participants than buying from a sole supplier. All the same, current trends favor collaborations of this kind, as can be seen in the auto parts industry. They will be successful chiefly if they bring manufacturers additional cost advantages.

Building trust

Optimizing the level of integration goes so deep into the internal structure of both manufacturer and supplier that a strong mutual dependency develops. This typically means that a manufacturer confines itself to a single supplier (that is, single sourcing). The successful companies we studied are well down this track. In over 50 percent of cases, they buy parts of a similar type from a sole supplier. Only just over 10 percent of their less successful competitors have ventured as far (Exhibit 2).

While less successful companies are wary of becoming dependent on a single supplier, their successful peers have realized that such an arrangement naturally also makes the supplier much more dependent on its customer. As the volume of business from one customer increases, so the supplier’s obligations toward that customer extend for much longer and include greater responsibility, such as setting annual rationalization targets for the life cycle of a specific machine.

In this context, the maxim that cooperation must be based on mutual trust cannot be taken too seriously. But such trust can be established only when the desired improvements yield advantages for both parties. Early commitment by each partner allows energies to be concentrated on getting the most out of the collaboration, rather than on staking claims and securing fallback positions.

Commitment to collaboration demands new procedures on the part of manufacturers. In particular, it means more demanding, less opportunistic definitions of what is required from suppliers; fast preselection of suppliers on the basis of a design competition; and a binding decision on the sole or main suppliers.

Manufacturers’ requirements are frequently formulated under the pressures of the moment and motivated by each function’s self-interest

Requirements of suppliers. Manufacturers’ requirements are frequently formulated under the pressures of the moment and motivated by each function’s self-interest. A manufacturer’s design department may want to cooperate with the most competent developer in the early stages, long before purchasing is even involved. The supplier that design prefers, however, may not be at all attractive from the cost point of view. Consequently, purchasing normally insists that other—cheaper—suppliers are taken on, at least as second suppliers.

In many cases, the product is so tightly geared to the first supplier by this time that no change is possible. Alternatively, purchasing may manage to find a cheaper supplier that can at least cover partial deliveries. When this happens, the original supplier starts to worry that a competitor will profit from its intellectual input; purchasing becomes unhappy at having to buy at least some of its parts from the "much too expensive" supplier; and the design department has difficulty motivating the most competent developer for the next project.

This vicious circle can be broken only if suppliers meet four key criteria:

1 Development knowhow. Purchasing will often suggest suppliers that are able to produce cheaply, but lacking in the necessary design capabilities. But since the greatest potential tends to lie in intelligent design-to-cost, such companies will not make the best partners. Even today, some manufacturers still believe they can do the development themselves or subcontract it to an engineering company, and then give the blueprint to a "garage company" for low-cost production.

Generally speaking, however, savings from such an approach exist only on paper, since by its nature it prevents total product/process optimization. The experiment all too often ends in the supplier redrawing the engineering company’s blueprint to make it into a production document. To avoid such inefficiency, suppliers—with the exception of a few commodity makers—must be required to have development knowhow.

2 Manufacturing knowhow and appropriate cost structure. It is clearly important for a supplier to have the capability to achieve the technological levels and volumes a manufacturer wants. But it is even more vital for the manufacturer to be sure, even before inviting entries for the design competition, that a supplier has the structural costs, logistics systems (on-time delivery), and quality standards to enable it to meet requirements.

3 Stability of the supplier as a business partner. Since a manufacturer will ultimately want to use a supplier as its only source, financially unsound companies with poor profitability and no highly capitalized parent to bear liability should be excluded.

4 Loyalty. Though this is not normally a problem, caution is needed when the supplier has close equity links with a major competitor, or if the manufacturer is not a very attractive customer for the supplier.

Design competition. Suppliers that meet these criteria should be given the opportunity to qualify for collaboration by taking part in a preselection design competition with clearly defined goals for performance and cost.

Each goal must be realistic, yet ambitious enough to render inadequate a reliance on traditional system architectures and production concepts. Suppliers should be expected to take new approaches and to submit proposals—in terms, for example, of level of specification or product complexity—that reflect and exploit the changed business conditions. The competition should enable the manufacturer to reduce the number of contenders to a manageable, but still sufficiently large, number—preferably three to five—with which it will have to discuss its requirements in detail.

This preselection process must not take too long. The best approach is to tackle the various systems in overlapping stages, limiting the time per system to three to four months. The result will be a choice of suppliers and a number of design proposals, binding agreements on price for the entire life of a product, and undertakings on the part of suppliers to rationalize continually during series production.

Single sourcing. Adopting single sourcing and introducing intensive, trust-based collaboration can normally produce cost savings of 30 to 40 percent. By contrast, playing off several suppliers against one another brings single-figure improvements at best—and those are rarely sustainable. Any emergencies associated with single sourcing—"Supplier threatened with bankruptcy"—can be largely avoided by following the selection process already described.

Trust has to be earned. A manufacturer should vet its prospective partners carefully. Properly exercised, trust makes big demands on both sides; on the other hand, it makes day-to-day business far easier. Much of the productivity deficit in a manufacturer’s indirect functions can be attributed to checking up on suppliers. From design, purchasing, and logistics through to quality assurance, an enormous amount of time can be spent in this way, generating hidden costs for the manufacturer. The successful machinery and component manufacturers we studied, by contrast, sometimes had quality control carried out by the supplier (Exhibit 3).

NEW PROCEDURES, NEW BEHAVIOR

Though the vision of the new cooperation between manufacturer and supplier is often shared by both their top managements, the day-to-day behavior of employees at all levels may be an obstacle to overcome.

An important condition for optimizing vertical integration is new skills and styles of behavior throughout the organization at both manufacturer and supplier. In excellent companies, this change in behavior is championed by top management, and disseminated to all staff by a competent project leader. Practical implementation should be kept simple. Instead of being issued with intricate instructions, managers and staff at both companies should have the opportunity to try out unfamiliar approaches in real projects.

If pilot projects encounter skepticism or resistance ("We can’t change anything in the current machine design"), they should be treated as on-the-job training and be granted an extended schedule—for example, up to the changeover to a completely new machine concept. It will often be possible to incorporate some of the new ideas developed into ongoing production. As well as cutting costs even before the new model goes into production, step-by-step introduction also limits risk during the changeover. If the company is prepared and its management is strong, change can usually be accomplished quickly.

An approach along these lines sends a powerful signal, making the need for and the extent of change clear to staff at manufacturer and supplier alike. Change must then be vigorously put into action in individual functions. This will work only if managers are able and willing to lead the process, and if the company practises targeted job rotation between functions—chiefly between purchasing and development, but also involving manufacturing. These are necessary conditions because the changes are often of a radical nature.

Development/design

There was a time when suppliers did not usually become involved until after the overall concept was fixed—after system interfaces had already been implicitly determined and specifications more or less defined. In the new cooperative relationship, however, the development function has to define the most important systems, and therefore also the appropriate level of integration, in agreement with other functions right from the start.

For the development function the change is fundamental: it means relinquishing the detailed design of subcontracted systems

This approach is actually in the development department’s own interest, even if it means having to fix the system architecture much earlier than before. It means that developers can concentrate their efforts on development where it adds most value, rather than dissipating it in firefighting exercises at a later stage. (In principle, this change would be needed even without closer involvement on the part of suppliers, but it is not current practice in the majority of development and design departments.) What is less to developers’ taste is the fact that they can no longer call in a supplier of their choice casually at short notice, but instead have to seek the agreement of a cross-functional group.

For the development function, then, the change is fundamental: it means relinquishing the detailed design of subcontracted systems. Logically, it should also mean buying prototype parts from the series supplier and having that supplier perform tests. Only in this way can a manufacturer optimize its total costs by introducing "lean management"; only in this way will a supplier realize that it bears full responsibility, and can no longer rely on being second-guessed by the manufacturer.

Purchasers will be judged by whether they can build up a detailed knowledge of their suppliers’ cost structures and performance capabilities

Purchasing

In the past, purchasing’s main contribution was often the annual negotiation with suppliers, in which the department tried to keep prices constant despite inflation. But the purchasers of the future will be judged by whether they are able to build up a detailed knowledge of their suppliers’ cost structures and performance capabilities, and use it when negotiating long-term contracts. Such knowledge will also make it possible to get the very best suppliers involved in the preselection competition stage. While today’s purchasing departments tend to be aware of most of the traditional suppliers, they have frequently not heard of those international players that have risen to world-class ranks.

The same background knowledge is also useful in contract negotiations, since it helps in understanding a supplier’s room for maneuver. Manufacturers that know a supplier’s costs and still leave it a profit margin are smarter players in the long run than those that try to squeeze out every last drop. Partnership is only possible when it is based on fairness.

Management control

At first sight, the control function might be expected to change least under the new regime. Often, however, it will have a tough struggle, since it has to abandon the long-prevailing marginal cost approach. It must switch to calculating the full cost of components and sub-systems early in development, and to working out how far fixed costs can be cut if parts are outsourced. To achieve truly sustainable cost cuts, it must aim for proportional reductions—higher ones would be even better—in the indirect functions, and drive through the corresponding processes of change.

In addition, control has to make detailed cashflow calculations to enable management to compare offers from external suppliers with the cost of in-house manufacture.

Production

By their nature, most decisions on reducing vertical integration will entail far-reaching changes in production. A company must decide which of the modules that are not essential for technological differentiation can still be handled competitively by in-house production. Its performance must satisfy the criteria of factor cost, economies of scale, productivity, and design-to-cost. In-house manufacturing may measure up to suppliers only if the production divisions concerned are spun off as profit centers, where possible with their own design and development departments—becoming, in a sense, suppliers themselves. These units should then concentrate on design-to-cost potential, and in the medium to long term on achieving structural cost advantages too—in particular, lower factor costs.

In-house manufacturing may measure up to suppliers only if the production divisions are spun off as profit centers

With vertical integration reduced, production must not only cut direct costs, but also trim its overhead to match. Ideally, the overhead should be reduced more than proportionately. In our experience, facility size is no advantage when it comes to infrastructure costs; rather, the related "administrative effort" increases with size. The only justification for large facilities is economies of scale in process technologies that are only cost-effective above a certain level of production.

Apart from that, "small is efficient" is practically a law of nature. Overhead costs should therefore fall steeply as vertical integration is gradually reduced. But that will be a possibility only if the infrastructure and manufacturing hierarchies are challenged aggressively.

"Small is efficient" is practically a law of nature

Organizational inertia must be overcome by the use of precise, tough targets derived from best practice benchmarks and extrapolation from a company’s current position. This is the only way to challenge and, if necessary, to change entire levels of hierarchy or supporting functions such as internal transport and maintenance. Restructuring such functions as logistics, say, is a simpler and more straightforward procedure than changing manufacturing and assembly. Such in-house processes as receipt of goods or transportation should therefore be carefully scrutinized.

Quality assurance

For this function, the new mode of collaboration with suppliers is a good opportunity to formulate the changes that need to be made in any case. Quality matters should be taken into account from the first stage of the supplier selection process, and a detailed quality audit should be conducted for the final selection. Quality control should also specify the quality targets for components and systems early in development.

With these changes, quality assurance will have less work to do in the production stage. Excellent companies are now relying much more on a supplier’s outgoing inspections and have reduced their own checks on incoming goods, but it has to be admitted that their level of trust in suppliers is still fairly low: just under 40 percent for machinery makers and a little over 20 percent for component manufacturers (Exhibit 3). There is clearly a need to work on building trust with suppliers.

Successful machinery and component manufacturers can reduce their total cost for a sub-system by up to 40 percent by optimizing their level of vertical integration. This process must cover the entire value-added chain, including the structures and processes of suppliers. It is the only way to exploit by far the most important opportunities for improvement: outsourcing of entire systems, design-to-cost to meet the needs of both partners, and consolidation in a small number of key suppliers.

Depending on the company, and above all on the strength of its management, the change process will take from two to five years. The chief stumbling block is lack of skills. With current moves toward making older and often more highly skilled employees redundant, this could pose a threat. Changes in behavior are essential at both manufacturer and supplier. What will ultimately make the difference is a committed management team with a strong will to change.