Working Paper 94-5

Demand Flow Technology® for Transnational Companies

Emmanuel A. Kampouris, President and Chief Executive Officer, American Standard Inc. Robert A. Miller, Associate Professor of Industrial Administration, Carnegie Mellon University

Demand Flow Technology®, or DFT for short, brings into focus a concept that companies competing on world markets, currently burdened with batch manufacturing processes, and managed by bureaucratic hierarchies, can only ignore at their peril: the sole rationale for firm expenditures...raw materials, labor, work-in-process inventory, administration...is to meet income generating demand. The concept of DFT is to pull raw materials and products through the process strictly according to the dictates of customer demand. Methodical, accurate analysis of the product and the process which yields it, including a detailed knowledge of the value added and nonvalue added steps, as well as the dimensions and standards of quality requirements, are a prerequisite for the successful implementation of DFT. Under DFT it would be inimical to produce final product or work-in-process on a just-in-case basis. Since the nature of demand and the state of technology are ever changing, a factory organized according to DFT principles is of necessity, flexible, with the ability to incorporate innovations as they occur. Whereas most innovations are small, requiring only minor adjustments to a plant floor already utilizing flow manufacturing, t he conversion to DFT requires major changes in corporate culture, organizational structures, management attitudes, worker relationships as well as process-line designs from a traditional plant. 

This paper explains how and why DFT is currently being implemented in American Standard, a $3.8 billion transnational corporation. It is illuminating to study this case for three reasons. First, it demonstrates the flexibility of one transnational in adopting completely new organizational forms to survive threatening global economic forces. Second, it illustrates why DFT in particular, is being embraced. Third, by tracking the states through which DFT is being introduced in American Standard, we can peer into the future as to how this concept might further evolve and what other changes might be expected of transnationals. 

The first section of our analysis sets forth the context for the conversion of American Standard to DFT. There we list the global factors which help explain why companies like this one have become transnationals, and how this, in turn, has led to the adoption of organizational forms that have previously been regarded as Japanese in origin. 

The adoption of DFT by American Standard, an employer of 35,000 workers at 82 plants in 34 countries throughout the world, is a mammoth undertaking, requiring top down unequivocal commitment to the technology as the plant's "way of life" rather than just another project. The adoption process is described in Section 2. Outside consultants were commissioned by American Standard to provide orientation courses and workshop training for employees from every aspect of the business and every facility within this global corporation. Additionally, the consultants are used to assist in actual implementation of DFT on the plants floor. After formal training in the technology, implementation teams are established at plant level, comprising a team leader and representatives from supervision, engineering, shop floor and union. At the manufacturing plant DFT is implemented in stages from final assembly back through raw material conversion...in that order. 

Section 2 also briefly discusses broader applications for DFT, in the office, as a simple yet appealing cost measure (which provides a basis for making product and pricing decisions), and at the research and development stage. 

Then, in Section 3, we review the progress of DFT at American Standard to date. Overall, it has been spectacular. In the period 1989 through 1992, product cycle time has been reduced by 90 percent, rework on defective products has declined by 70 percent, working space has halved, labor productivity has increased by 20 percent and inventory turnover has almost doubled. Comparing the achievements of individual plants, we see that, broadly speaking, the gains from implementing DFT accrue over a prolonged period, larger plants benefit from DFT more than small ones, and that cultural considerations seem to be more important before DFT is introduced than afterwards. 

These latter findings lead us to conclude, in Section 4, that while culture may play an important role on inefficiently organized shop floors, and while the conversion costs depend on cultural norms, the organizing principles of DFT are largely culture free, and consequently an unparalleled degree of uniformity across plants throughout the world is achieved by implementing DFT, a factor which greatly strengthens the positions of transnationals in the global marketplace. 

Take the second factor. Protectionism has been retreating for some time now, and foreign currency exchange markets are becoming more sophisticated. Over the years GATT has made significant inroads almost across the board in terms of products and countries. The Common Market, the EEC, and now the EC, represent epochs of ever more liberal trade regulations within Western Europe. On a smaller scale, the free trade zone being created in North America is also removing trade barriers. Finally we believe the collapse of the Soviet empire will ultimately facilitate trade with Eastern Europe, if not Russia. 

The relevance of the first two factors is obvious, but the ramifications of the last one warrent amplification. The percentage of the world GDP produced in the U.S. fell from 40 percent in 1965 to 30 percent in 1987, the decline being almost exactly offset by Japanese growth, the total share of the 3 other largest economies, Germany, France and U.K. holding constant. In relative terms U.S. consumption has not fallen as quickly, and the shortfall has been met with increased foreign investment into that country; for example new international investment in manufacturing by U.S. companies abroad has been substantially less than the amount undertaken by foreign companies in the U.S.which on average is (several) years older in the States. But equipment is only one aspect of production; how work is assigned, scheduling procedures, quality control, are some others. On these other issues the concensus is that, compared to Americans and Europeans, the Japanese have: 
(1.) simpler inventory control methods 
(2.) shorter production lines 
(3.) equipment with lower setup costs 
(4.) less inventory 
(5.) workers able to perform different tasks 
(6.) workers hours are more according to employer's discretion 
(7.) less floor space 
(8.) fewer defects 

How has this come about? Part of the reason is easily understood. A person invests early in their life (by going to school and later learning more job-specific skills) than later on: their consumption pattern follows an almost opposite trend, especially when you take into account their leisure time. Similarly, countries devastated by war but with a well defined system of property rights left intact (an important qualification) immediately invest a lot, and as they mature this tapers off, consumption growing as their assets pay dividends. The relative positions of the U.S., Europe and Japan in 1965 thus set the stage for the patterns of global investment we have seen. 

The interpretation of these trends also explains the perception that Japan and to a lesser extent Korea are calling the shots. Since they started with a lower base, and have invested at a greater rate, their capital equipment is newer. To the extent that there are complimentarities between capital of adjacent ages in design, one might expect them to build capital which is directed more to their own demands, taking the rest of the world along with them. This process of modernization is thus less difficult for firms which are representative of growth firms in Japan, but potentially quite disruptive for others. 

These changes operated through both the product market and the factor market. Japanese consumers (and voters) were willing to forego higher living standards 20 years ago in order to improve their lot today. Japanese managers get paid much less then their counterparts in America and Europe. Also, while Japanese workers are now paid as much as their American equivalents in terms of purchasing power of internationally traded goods, this is a relatively recent phenomenon, and furthermore higher housing prices and overall congestion within Japan still leave many Japanese workers worse off than their American counterparts. In other words, the business groups in Japan and the tiger economies may have operated as monopsonists on the labor market, extracting much of the surplus that comes from development. Western firms charged lower prices, paid higher wages and offered their employees better working conditions than their Far Eastern competitors. Because the profits of the latter groups were invested, it would be most surprising if Japanese firms had not become leaders in manufacturing technology and organization. 

(1) Air Conditioning Products. . . 

which employs 13,700 people at 24 manufacturing plants in 7 countries, accounting for 50 percent of the company's sales; 15 percent of sales from this segment come from outside the USA. <.P> 

(2) Plumbing Products. . . 
. . 

which employs 16,000 people at 45 manufacturing plants in 20 countries, accounting for 30 percent of the company's sales; 77 percent of sales from this segment come from outside the USA. 

(3) Transportation Products. . . 
. 

which employs 5,500 people at 12 facilities in 13 countries, accounting for 20 percent of the company's sales, all of which is generated outside the USA. 

Incorporated in 1929, American Standard traces its roots back to last century. During the late 60's early 70's the company participated extensively in the wave of mergers and takeovers, acquiring its transportation products sector (which markets under the WABCO trademark), and later the air-conditioning division (whose products are mainly marketed under the TRANE name). In 1988, the company became the target of a takeover bid by Black & Decker, provoking a management led leveraged buyout for $3.2 billion. Since then the company has defined a set of core businesses (which we just described), built upon those, and disposed of divisions which do not belong to this core. Between 1988 and 1991 5 non-core businesses representing $800 million in sales were sold for a total of $650 million. 

The challenges facing American Standard in the global marketplace are similar to those facing other transnationals. The Transportation Products sector sells air brakes and related systems to several commercial vehicle manufacturers; consequently the capacity and location of these plants is essentially determined by the vehicle manufacturers themselves. The reason why this sector has operations in 13 countries, then, is that the vehicle manufacturers served by American Standard are themselves multinational companies with plants all over the world. Although the plumbing sector sells a sizeable proportion of their output through wholesalers (rather than building contractors and retailers) and the air conditioning sector has a bias toward building contractors and installers, the location and demand of the specifiers themselves ultimately prescribe the geographic market for these products. Therefore the wholesale outlets or installers should be considered as part of the distribution system, complementary with distribution centers run by the company itself for example. 

When a new geographic market appears in these two sectors, a number of factors determined the firm's response. If the market is in the same country (or free trade/common currency zone) as an existing plant, then the firms must decide whether the net reduction in transportation costs (if any) from locating near the new market justifies the additional overhead incurred from setting up a new plant. (There may be labor relations issues and state government questions to consider as well.) There are further issues to consider when the market is in another country. On the one hand labor market and regulatory questions markedly change across national boundaries; on the other if the firm does not build a new facility but exports instead it opens itself up to the vicissitudes of trade policy (including tariffs, quotas, restraints and so forth) and greater exchange rate risk. To see the latter point, notice that under the export option, all sales revenue is subject to exchange rate risk, whereas if the new line is located within the host country only cash flows back to the country where shareholders reside is subject to this risk (that is, assuming liens are also taken out in the host country, or fully hedged). There is one further benefit from locating near the new market which, though we have saved to last, should not be understated. Proximity to the market helps the firm to learn more about the demand it faces and respond appropriately; this is particularly important if there is reason to believe the characteristics of the new market differ from existing ones. 

In American Standard's case many of these factors have worked towards establish- ing new plants. For example, Malaysia and Thailand currently demand a less sophisticated air conditioning product than the product sold in Taiwan, and this provided a rationale for setting up more than one plant in the region. In Indonesia another plumbing firm was taken over thus allowing the company to tap into an established market; low labor costs, isolation from other plants and steadily growing demand made this move look very promising. American Standard has had a similar experience in Czechoslovakia and Bulgaria: as capitalism is established there, a highly skilled labor force, and anticipated growth, have created an attractive investment opportunity in the plumbing products sector. 

Under DFT it would be inimical to produce final product or work-in-process on a just-in-case basis. Inventory becomes an impediment, and as a result consider- able reductions are brought about in raw material, work-in-process and finished goods inventories. This latter point was of utmost importance to the company, which found itself in a highly levered position and saw the reduction in inventory as a means to reduce its debt quickly. Since the nature of demand and the state of technology are ever changing, a factory organized according to DFT principles has to be very flexible, incorporating innovations as they occur. Whereas most innovations are small requiring only minor adjustments to a plant floor already organized along DFT lines, the initial adoption of DFT requires major changes. 
The institutional changes wrought by DFT in American Standard were imposed on the organization by a small cadre of senior managers following a management orchestrated LBO to avoid a takeover offer by Black & Decker; these managers committed American Standard to DFT as a that only those functions which contribute value to the firm should be retained. In times of economic hardship, it forces a company to be introspective, discarding those operations which are not demonstrably contributing to the whole. The resulting reorganization and realignment of the company is designed to make the company better prepared to meet challenges that the marketplace will offer in the future. DFT is implemented in several stages starting from the point at which product is ready for shipping and then moving back through the process. 

Describing the induction as a training program would be a gross misnomer. What each person was exposed to depended on their position within the company; executives had to gain an understanding of how the principles of DFT worked, ways of monitoring its implementation, how it would affect interactions between sister plants and so on, while plant employees were more oriented towards the effects of DFT on their particular part of the factory floor. 

Because DFT has far reaching implications which might appear unrelated to those employers who do not understand it fully, catchy phrases are frequently repeated in the company's monthly newsletter, awards are made to people doing an outstanding job of implementing DFT, and other psychological motivators are used to galvanize the workforce and impart a sense of overall purpose. For example, every American Standard employee knows that the letters "TNT" stand for "Twice the Turns Now", a slogan for reducing inventory by better synchronizing the manufacturing flow process. These letters appear all over the place, even inscribed on chocolates and ties given to company guests....conversation pieces to drive home the point. 

The implementation teams provided more informal, yet just as powerful, guidance as well. An anecdote illustrates this point. While they were interviewing a team of workers on the job just after adopting DFT on one assembly line, an implemen- tation team member suddenly blurted out in feigned blank astonishment: "Look we've stopped the line, just by talking to you (as if this was the cardinal sin). If you're being spoken to, step back out of the line and allow your team workers to cover your position instead of bringing everything to a standstill." This single event became a lesson, seeping all the way back to the plant manager. 

The use of informal motivators has been adopted by the company itself. In Brazil, the company controller was asked what would happen if they had a shortage of workers in the plant. Without batting an eyelid he said: "I would send my office people to fill the void". Needless to say his response had quite an impact, epitomizing worker flexibility. Again more than one person learned that day. Another example: Kampouris, the CEO, literally takes snapshots of the work-in-process as he listens to plant managers promise how much it will be reduced in the coming months. Partly because they are unanticipated and partly because they have not been institutionalized, clever informal motivators like these have been effective in convincing the workforce of American Standard that DFT is not really a slogan or set of work rules and accounting procedures, but must become a "way of life" at the plant. 

Thus, despite the fact that some fully automated subprocesses were replaced by less automated processes (in the interests of reducing work-in-process, inventory and setup times) the output flow of defect free products actually increased and inventory turns improved. This is reflected in cash flows as ordering for components is temporarily slowed to use up "excess" parts already in stock and manufacturing for stock is eliminated). Over time the worker teams acquire broader experience about manufacturing the product. The most visible way in which this knowledge is institutionalized is through "method sheets" which provide colored graphical instructions on precisely what must be done at each operation, including TQC and verification checks for any given product being assembled or produced on that line. 
This acquired experience is also incorporated into the product design changes. In America there is no precedent for factory operatives explaining to the design engineers what their requirements are for better manufacturability. This illustrates the flip side of accountability: "empowerment". 

With regards inventory, turnover has almost doubled since 1989. Putting this another way, the value of DFT can be roughly measured by the lower costs of inputs generated by reducing inventory. By 1992 the accumulated 3 year savings in inventory had accounted for roughly 10 percent of annual sales! This leaner organization is reflected in the decline of working capital (as a percentage of sales) from 16.5 percent in 1989 to 8 percent in 1992. The bottom line is that over the past 3 years cash flow has substantially exceeded sales less costs. 

All three sectors, plumbing, air conditioning, and automotive, have contributed substantially to the gains DFT has wrought. For example, inventory has fallen 25 percent over the last 2 years in the automotive sector, and by more than that in many plants in the air conditioning and plumbing sectors. On other dimensions, such as product cycle time, customer lead time, space savings, and employment the improvements registered across the 3 sectors, is the same order of magnitude. Variation in gains across plants within a sector is much more pronounced than variation of the plant average across sectors. These facts support our view that the gains from DFT are not systematically related in a simple way to the machines or engineering technologies used in production. They come, primarily, from new organizational and human capital, which achieve stronger symbiotic relationships between workers, machines and the products they manufacture. 

The second point to note is that, for a given implementation date, bigger plants achieved greater inventory reductions than smaller ones. This is certainly true in absolute magnitudes (which is the relevant measure for profits), but the percentage reductions tend to be more impressive at larger plants as well. There is another, more subtle, piece of evidence in the table that suggests, more generally, that bigger plants benefit more from implementing DFT than smaller ones. So if American Standard anticipated this they would have left their smaller plants until last. For the most part, this is what happened within each country. For example in England the largest plant implemented DFT in 91-1, the next largest in 91-3 and the smallest (shown in Table 1) in 92-3. 
 
The literature on organization suggests several reasons why larger plants have more to gain from adopting DFT. Large organizations are more unwieldy to manage, partly because there is more potential for disfunction, and also because it is harder to enforce an incentive structure that serves the organization's goals. DFT provides a set of rules, or norms, that sort out job priorities in a fairly straightforward manner. This is one reason explaining the difference in the results between large and small plants. In addition, larger plants enjoy the scale advantages of being able to better balance the production lines and thus can better capitalize on DFT. This factor may prompt some consolidation of plants if output at the smaller ones does not grow significantly in the future (and indeed this is already taking place to a limited extent). 

Several other features are evident. The sample variance of both the customer lead time and the product cycle time have fallen quite substantially. As we have mentioned before, DFT imposes a discipline on the factory floor which channels all efforts to meeting domestic demand subject to the constraints implied by the production technology. Roughly speaking, a common production technology is shared across plants in different parts of the world within this company. (For example, plant managers have many opportunities to review the production process at sister plants elsewhere, and high level executives are rotated throughout the world as a matter of course.) In contrast, demand from each of the plants and the supply of raw materials is largely regional and therefore subject to the vagaries of local markets. Thus we are not arguing that the plant performance across the world be equalized on any of these dimensions. Indeed, differences in wages and other factors imply that some variation will be induced by adapting DFT to each specific environment. What the relatively small variances amongst plants adopting DFT means, is that technological considerations associated with DFT outweigh those arising from differences in local demand and supply patterns. 

Coupled with their high mean times, the high variances of pre-DFT days recall to mind the saying, that there are many wrong ways of doing the job (but only one right way). However there are no obvious patterns that emerge from the plant data which establish that their previous inefficient production methods were culturally related. We note that DFT tended to be introduced first in the USA, then in Europe and last in Asia. One reason for this brings us full circle to the beginning of our talk: since plants in the US are now amongst the oldest in the developed world perhaps it is not surprising to see renewal beginning there. But cultural factors also played a role; since the language used by company executives is English, there were fewer communication hurdles to overcome by introducing DFT first to plants where English is the lingua franca. To summarize, while DFT is largely free of cultural norms, the benefits from implementing it vary across the world, because previous management practices in some countries conform more closely to DFT than in others, and learning about DFT may be easier in some countries than in others. 

The literature on international management is replete with examples of multinational companies failing to adapt to host investment companies, by enforcing inappropriate organizational norms with managers who are culturally bound to the nationality of the central office postal address. These examples are sometimes used to motivate a discussion of tensions between different management cultures and how they can be ameliorated in multinationals. Quite frankly, these discussions are a little dated. We are not arguing that many firms wrestle with trade-offs between tensions created by different cultural backgrounds, merely that such compromises are inefficient and unlikely to survive global competition very far into the next century. As we have demonstrated, DFT is supported by pictures, charts and numbers that are largely free of cultural norms. Job assignments and tasks can be molded to the plant machinery as carefully as is humanly possible. The overriding concerns in this new-age factory architecture are related to the demand flow for products, the machinery used in production and the motor skills or physical dexterity of plant operatives. In this way differences in cultural backgrounds are essentially removed from the plant floor altogether.