LEAN MANUFACTURING – DESIGNING A CAR – The Techniques of Lean Design

It’s wonderful to know that new products can be produced faster with less effort and fewer errors. However, innovations are most useful when they are available to everyone and practices at General Motors, Volkswagen and the other mass producers, lag far behind. I know because myself I was doing this for years and during the product development work on projects for different car OEMs I had the opportunity to see how things are done on both sides as mass-production apporach and also as lean-production. Therefore the following question must be adressed:

What then are the precise techniques of lean design that the best auto firms use and how can they be transferred to existing mass producers?

From what I have seen during my professional career as design engineer, comparing the 2 major players in car manufacturing industry I was working with, namely Toyota and Volkswagen I can only conclude that there are at least 4 basic differences in design methods employed by mass (such as Volkswagen) and lean (such as Toyota) producers. These are differences in:

• leadership
• teamwork
• communication and
• simultaneous development.

Taken together, lean techniques in these 4 areas make it possible to do a better job faster with less effort. But let’s have a closer look at each of them.

LEADERSHIP

1st, let’s look at the leadership of a project.

The lean system incorporates cross-functional collaboration at the beginning of a new project not at the end like applied in mass production systems. The lean system comes with more formal authority to enforce accountability.

The lean producers (mostly Japanese) invariably employ some variant of the shusa system pioneered by Toyota (also termed the “large-project leader”, or LPL, system at Honda). The shusa brings together a small team from different functional departments that work on the project for its lifetime. Team members would sign formal pledges detailing the work that they committed to doing. By using this approach, the lean team puts the most time and effort determining resource allocation and trade-offs at the beginning of the project, but once a state of consensus is reached, fewer and fewer people are required to be involved later in the process.

The shusa is simply the boss, the leader of the team whose job it is to design and engineer a new product and get it fully into production. In the best Japanese companies the position of shusa carries great power and is, perhaps, the most coveted in the company. True, employees may seek the position as a stepping-stone to the top. However, for those who truly love to make things, the job brings extraordinary satisfaction. In fact, it’s the best position in the modern world from which to orchestrate all the skills needed to make a wonderfully complex manufactured product, such as the automobile, coming into being. One might even say that shusa is the new supercrafsman, directing a process that now requires far too many skills for any one person to master. Oddly, while we’re used to thinking of dedicated teamwork as the ultimate sublimation of individuality, new products inside the Japanese auto industry are commonly known by the shusa’s name: “That is Fuji-san’s car” or “Akoika-san has really stamped his personality on that car” are phrases commonly heard within Japanese companies. Perhaps after all, we cannot escape a human need for craftsmen to exist. However, in an era when the skills involved are not so much technical as social and organisational – and far beyond the grasp of any individual – craftsmen must now take the form of the shusa.

The Western mass producers also have development team leaders, as for instance Volkswagen or GM. What’s the difference between the two systems? I believe it lies in the power and career path of the team leader.

In Western teams, the leader is more properly called a coordinator, whose job it is to convince team members to cooperate. It’s frustrating role, because the leader really has limited authority, so few team leaders report enjoying it. Indeed, many company executivs view the job as a dead end in which success leads to little reward and failure is highly visible. (I don’t have a more evident example of that, than what Volkswagen did some years ago with DieselGate story, which is the obvious result of such practices).

What’s more, the team leader is in an extremely weak position to champion a project within the company. It’s common in Detroit (at GM) , Wolfsburg (at Volkswagen), Ingolstadt (at Audi) and Paris (at Peugeot) for top management to override the team leader about the specification and feel of the product – often repeatedly during the course of development. That this happens is understandable, given senior management’s role of juggling other corporate needs as market conditions change. However, in the worst case – and all too frequently, particularly in the United States – the result is a product with no personality or distinction that the company must sell solely on the basis of low price.

TEAMWORK

The problem here becomes clearer by looking at the second element of lean design, the tightly knit team.

As I’ve mentioned earlier, in the lean-development process (for instance at Toyota), the shusa assembles a small team, which is then assigned to a development project for its life. These employeed come from functional departments of the company – market assessment, product planning, styling, advanced engineering, detail engineering (body, engine, transmission, electrical), production engineering and factory operations. They retain ties to their functional departments – it’s vital that they do so – but for the life of the program they are clearly under the control of shusa. How they perform in the team, as judged by the shusa, will control their next assignment, which will probably be another development team.

By contrast, in most Western companies (such as Volkwagen & GM), in mass productions systems a development project consist of individuals, including the team leader, who are on short-term loan from a functional department. Moreover, the project itself is moved from department to deparment along a sort of production line, which leads from one end of the company to the other: that is , the project is actually picked up and moved from the marketing department to the engineering divisions, and then to the factory operations department during its life, in the same way that a car moves from the welding to the painting to the assembly department in the asembly plant. So it is worked on by totally different people in each area. In this case the members of the team know that their career sucess depends on moving up through their functional specialty – getting promoted from chief piston engineer to deputy chief engine engineer to chief engineer, for example – and they work very hard in the team to advance the interest of their department.

In other words, being a member of the GM-10 team as it was the case in 1980s or even of the more recent Audi E-Tron team, say, doesn’t lead anywhere. The team leader will never see an employee’s personnel records, and the leader’s performance evaluation won’t make much difference to the employee’s career. Key evaluations will come from the head of the employee’s functional division, who wants to know, “What did you do for my department?” As a result, discussing the best way to achive harmony between the engine and the body, for example, easily can disintegrate into a politicized debate between the interests of the engine engineering department and the body engineering department.

The continuity in the Japanese development teams is reflected by the fact that about for instance about 900 engineers are involved in a typical project in an American (GM) or European (VW) company over its life, while a typical Japanese team enlists only about 485. The Lean system could move faster, stay on schedule, and offer a wider variety of products as a result.

What is more, those Japanese firms most committed to the shusa system (a.k.a.“heavyweight” team management) needed an average of only 350 team members, while the Western firms with the weakest teams (mostly in Germany at Audi and VW) needed an average of 1.500 staff members during the life of a project.

The Japanese use fewer people partly because efficient organization requires fewer bodies, but also because there is so little turnover in the Japanese teams. Because Western department managers view team members as simply the representatives of their home department in the development process, they show little concern about frequently recalling staff as other needs for their skills crop up in their own department. For the team, however, these recalls mean a great loss because much of the essential knowledge of a development team lies in the shared viewpoints and experiences of team members over an extended period.

COMMUNICATION

Many Western development efforts fail to resolve critical design trade-offs until very late in the project. On reason is that for instance the U.S. team members show great reluctance to confront conflicts directly. They make vague commitments to a set of design decisons- agreeing, that is, to try to do something as long as no reson crops up not to.

In Japan, by contrast, team members sign formal pledges to do exactly what everyone has agreed upon as a group. So conflicts about resources and priorities occur at the beginning rather than at the end of the process. Another reason is that a design process that is sequential, going from one department to the next rather that being kept at team headquarters, makes communication to solve problems very difficult in any case.

The result is a striking difference in the timing of the effort devoted to a project. In the best Japanese lean projects, the numbers of people involved are highest at the very outset. All the relevant specialities are present, and the shusa‘s job is to force the group to confront all the difficult trade-offs they’ll have to make to agree on the project. As development proceeds, the number of people involved drops as some specialities, such as market assessment and product planning, are no longer needed.

By contrast, in many mass-production design exercises, the number of people involved is very small at the outset but grows to peak very close to the time of launch, as hundreds or even thousands of extra bodies are brought in to resolve problemd that should have been cleared up in the beginning. The process is very similar to what I saw in the assembly plant: The mass producer keeps the line moving at all costs but ends up doing massive amounts of rework at the end, while the lean producer spends more effort up front correcting problems before they multiply and ends up with much less total effort and higher quality in the end.

Mass production is the image you might conjure when thinking about a classic assembly line. Individuals stationed along a conveyor belt with a single task to complete, enabling a faster pace of production and jobs for low-skilled laborers. But in this model, work is usually precarious, as workers are easily replaceable, and the worker doesn’t have say in how systems can be improved. Factory managers were pressured to keep to their production targets, therefore set the expectations that the line should never be stopped unless absolutely necessary. As a result, if parts at your station aren’t fitting correctly, you will probably send it forward in production to keep the line moving. This all results in a fast rate of production, but a huge backlog of issues with the fully assembled vehicles, which requires time-consuming troubleshooting and rebuilding before the vehicle can be finished.

When Taiichi Ohno was developing his model for the ideal factory floor, he was faced with a very important constraint: a labor settlement that required lifetime employement for the workers. Within this context of labor as a fixed cost, he saw an opportunity to utilize this human capital to the fullest extent. So in practice workers were given more complicated work and a role in improving the production prosess. They were given access to information using digital disply systems, so they could respond quickly to problems. Teams would go through the process of asking “the 5 Whys” getting to the root of problems and implementing effective solutions. Famously, every worker was also given the power to stop the line. Once the line was stopped, teams would start troubleshooting the issue and tracing its origins. These discussions might be seem to an outsider as inhibiting their efficiency, where in fact, the brief investment in problem-solving can prevent much larger and more time consuming issues from occurring in the future.

When comparing two comparable plants (for instace Volkswagen vs. Toyota, or GM vs Honda) in 1987, the lean plant was almost 2x more productive and 3x more accurate, using 40% less space. This is very much the case today as well, that’s why Toyota has become the most succesfull car manufacturer in the world and they are keep doing this for quite a a long time already, without significant signs of slowing down or loosing market share.

SIMULTANEOUS DEVELOPMENT

Simultaneous Product Development, or SPD, is a method of design collaboration in which members of the entire global development chain work in parallel, synchronously or asynchronously, to create and finalize product definitions. Simultaneous means at the same time. Companies that practice Simultaneous Product Development achieve the closest possible collaboration between designers and the rest of the development chain.

Simultaneous product development is based on the premise that specialists from every area should be able to directly contribute to product design as early and as frequently as possible. These experts should also have the ability to interact with any detail of the product. They should have the tools to reconcile the product into the final specifications, and keep a record of the several paths and variants.

So the final technique separating lean from mass production in product development is Simultaneous Development. To see what I mean by this term, let’s take the example of die development.

Practically every car and light truck build in the world today has a body constructed of stamped steel panels. The heavy metal forms, called dies, needed to press finished body panels out of sheet steel are among the most complex and expensive tools in the industrial world. They are made of exotic steel alloys for extreme strength and hardness and must be formed to tolerances of microns across continuously curving surfaces. What’s more the matching faces of the die (the upper and lower or “male” and “female” elements) must mesh with absolute precision. Otherwise, the sheet steel will tear or even melt to the face of the dies as the 2 pieces come together under tons of pressure.

The mass-production approach to die-making has been simple: Wait until the product designers give precise specifications for the stamped part. Then, order an appropiate block of steel in the die-production department and cut it, using expensive, computer-driven, die-cutting machines. Because of the cutting proceeds through many steps involving many machines, this process means that the dies pile up waiting for the next machine to become available. Total development time, from the first day that product designers order a new set of dies until the dies begin stamping panels for production cars, is about 2 years. (this appoach is commonly followed at Volkswagen and Audi)

By contrast, the best lean producers – and they’re all Japanese but no longer only in Japan (for instance Honda is designing and cutting dies for its Marysville, Ohio, plant at Marysville) – begin die production at the sale time they start body design. How can they? Becasue the die designers and the body designers are in direct face-to- face contact and probably have worked together in previous product-development teams. The die designers know the approximate size of the new car and the approximate number of panels so they go ahead and order blockes of die steel. Then they begin to make rough cuts in steel , so it’s ready to move to final cutting as soon as the final panel designs are released. This process of course, involves a considerable degree of anticipation. The die designer must understand the panel-design process as well as the panel designer does and be able to anticipate accuratelly the panel designer’s final solution. When the die designer is correct, development time is drastically shortened. When the die designer is wrong (an infrequent occurrence), the company pays a cost penalty. Still, the original schedule can be met by giving the catch-up die priority routing through the cutting process. Also, the lean die makers seem to be much better at scheduling production in the die-cutting shop. Their solution should come as no surprise: the die cutters have special, quick-changes cutting tools, allowing one machine to handle many different types of cuts, so the dies that are being cut spend much less time in queues.

What’s the end result of this intense communication between panel designers and die makers plus accurate anticipation by the die makers and clever scheduling of flexible cutting machines? It means that the best lean producers in Japan (and in Ohio) can produce a complete set of production-ready dies for a new car in 1 year, exactly half the time needed in typical mass-production die-making. Not surprisingly, this process requires fewer tools, lower inventories (since the key element, the expensive die steel, is in the shop only half as long, ) and less human effort.