PART 1 of 2
Assembly productivity is vitally important to car companies, although its relative importance in the total car cost picture is declining, as factories become more efficient worldwide, and as outsourcing of more parts and assemblies to suppliers continues. Nevertheless, factory productivity remains a key competitive metric for all car producers.
There is general agreement that some variant of "lean production" or the Toyota Production System (TPS) is the way to go in terms of overall manufacturing strategy, but implementations of TPS vary widely, by country, by company, and even by plant. There is significant difference in basic productivity measures (e.g., hours of labor required to build a car) even across various plants that were each considered leaders in the industry, despite adjusting for vertical integration differences, model complexity, etc.
Here we'll share some of the general sound practices that cut across most plants, based on numerous interactions with assembly plants in recent years and in many countries. Our focus is on factors plant managers can work on, such as process parameters and operational decisions—and less on "upstream" decisions such as plant or product design, which, of course, also have major impact, but which cannot be altered in the short term. (Consequently, for example, we'll omit a discussion of line-side parts delivery—as good an idea as it is—because it involves strategic decisions about sourcing well above the plant manager level.)
In this article we'll provide an overview of recommended practices at the total plant level. In Part II (AD&P, September 1998) there will be examples of good management at the plant sector or "shop" level (including stamping, welding, painting, and assembly shops), with some final comments on the support services that serve all the shops. We'll avoid using too many specific numbers, as these can be easily taken out of context (for example, hours per car can be much higher in a low-wage-environment country, without any economic penalty).
Two general caveats about our observations: first, any quantitative measures we discuss are based on a "standard car," essentially a mid-sized four-door sedan as is sold in the European C range. Without such a standard, comparison across factories become meaningless, as one would not expect the same amount of labor to go into a luxury four-door sedan as into a two-door "econobox." Adjustments must therefore be made to the numbers if the reader wishes to draw conclusions about a car significantly different from this standard. Our qualitative recommendations, such as about the use of teams, apply to any type of car.
Secondly, we omit in our discussion any radical new manufacturing systems such as ultra-modular plants like the one used for the Smart car. By definition, optimized but still traditional "lean" approaches may not apply to these plants. Our work focuses on how an assembler might reach world-class levels within the mainstream manufacturing paradigm, not on how one might invent a new paradigm.
Overall Good Practices at the Plant Level
Successful plants distinguish themselves at the overall factory level along seven key dimensions: team structure, cross-functional empowerment, employee motivation, workforce competence, services outsourcing, short-term production management, and long-term production planning. Taking each in turn, we find the following characteristics:
•Teams. Car plants really can't run without teams, due both to the complexity of the work and, ironically, its monotony, which demands job rotation during the day. Thus, a team must be formed to work on some part of the line. It needs to be sized in order to comprehend some significant potion of the work, and to maintain rotation and learning opportunities. Yet it can't be so large as to lose "team spirit." Varying with the work organization, we found effective teams of 10 to 25 people. The team leader job is the most pivotal in the whole plant, and thus demands the most careful attention. We found team leaders most productive when they were fully plugged into the team rather than riding above it. This means each team leader needs to know every team member's job, be able to competently carry it out and teach it, and be well connected with the leadership of adjacent teams. Good team leaders are actually working on the line at least half of the time. This gives them great task insight and builds their credibility with the team members. Finally, it is critical that the team leader receive regular performance reviews that include target-setting discussions specifically focusing on continuous improvement in line balancing, defect reduction, and downtime minimization.
•Cross-functional empowerment. No team can function effectively without having at its disposal all the tools it needs to correct problems on the line. After all, a plant has two roles: efficient production of the current model, and continuous improvement of the process. The key factor for success here is to ensure that line workers are not only trained to do their specific jobs, but that they also have both the training and the authority to perform minor maintenance and QC associated with their job. Otherwise, it is very hard for the factory to improve. This requires flexible job designations (so that the maintenance function doesn't become an unresponsive "empire"), supportive incentive schemes, and extensive line worker training in both repair and quality control techniques. The ideal, of course, is to enlist workers in continuous job redesign and improvement as is done at Toyota and other companies.
•Worker motivation. The car assembly task is so complex that no manager can possibly handle all the direct commands that would be necessary to build a car with a hostile or ill-prepared workforce. Beyond just good "team spirit" and cross-functional task empowerment, there are myriads of routes to high worker motivation—none especially easy—but we find the best plants both instill motivation, and capture its best effects, by running very comprehensive and effective suggestion systems. Honda of America Manufacturing's system is one of the very best. A significant majority of the workforce should be contributing to the system each year, and each worker contributing should produce a high number of suggestions. This is both to encourage continuous improvement and to stimulate creativity. To avoid demoralizing the workers, a very large percentage of submitted suggestions should be implemented: many plants aim for 80% or more. In terms of measuring motivation, the one most vital statistic in this area is unexcused absenteeism, held to no more than 2% if at all possible (recognizing there are cultural variations). Not only is very low unplanned absenteeism destructive to lean production (which by definition runs with no surplus workers standing around ready to take up the slack), but many plant managers have found a worker's attitude toward absenteeism to be the single best predictor of that worker's long-term potential and motivation. That is, a willingness to show up predictably and put in a good day's work seems to be more important than previous work experience, training or skills. (One can conclude that one reason the Japanese transplants in North America chose rural sites was to tap into the agricultural work force, which is well-used to good attendance, as the crops wait for no one!)
•Workforce skill. Of course, a solid level of workforce competence is necessary as well as motivation and attitude—though perhaps not as important: tasks can be learned, enthusiasm cannot. Skill building begins with recruitment screening, which seems most effective when everyone gets into the assessment process: management, the union, team members, and the assessment center staff themselves. Commitment to finding the best candidates grows when more parties have a stake in the outcome. Once hired, integration into the team-based working environment and job training become the next critical steps. The most effective training is split in its delivery mode: technical job training must be delivered on the line by a respected and competent operator (not a trainer) whileenabling training—such as quality control techniques and people skills development—can be provided offline, typically at the company's training center.
•Plant services. Any services that don't directly add value to the car should be outsourced, plain and simple. This allows the teams to focus on the key tasks of building the car, and learning to build the car better. From the local college students who give the plant tours to the suppliers performing the material handling on their own deliveries, all non-production tasks should be ruthlessly peeled away and given to someone else. The rationale for this is manifold, from issues of morale (no regular employees should be saddled with "second class" work, nor allowed to hide out in easy support functions), to reasons of efficiency (workers are freed of any burdens save the need to do their specific tasks well, and then to strive to improve them).
•Production management. In the area of short-term production operations, perhaps the key lessons are to plan extensively and to build up flexibility. In terms of planning, the focus is on line balancing and re-balancing, aiming for zero balance loss. This goal is certainly unobtainable, as it is a moving target, but the point is to pursue continuous improvement. No modern plant should ever consider its lines "done" in terms of balance, but rather should expect a myriad of micro-adjustments in the line, even from week to week. Such fine-tuning is, perhaps, unrealistic, especially in modern low-inventory systems, as it assumes a constant high level of resource availability, and "no mistakes." Both of these assumptions are unreasonable, and so an advanced car assembly plant has to build in enough flexibility to handle disruptions and shortfalls. The two key practices in terms of flexibility are delegation of process time fine-tuning to the teams on the line (requiring sound communication between the teams and the production engineers, who must let go of some of their prerogatives), and daily floating overtime availability. The latter is tough to put in place in certain national work environments, but is critical to success. It also leads to more secure long-term employment, even if at the expense of short-term employee stress. Many advanced plants (especially in Japan), in effect, deliberately under-staff by 5% or so, preferring periodic bursts of overtime for the core of the workforce to periodic layoffs of "peripheral" workers.
•Production planning. Finally, the best plants are distinctive in terms of longer-term production planning. The objective is to provide as much stability in the mid-term as possible, so that the short-term flexibility just discussed is not overloaded. Practices vary somewhat, but two are quite common. First, there must be an absolute insistence on production process stability. Each new process must be thoroughly debugged and brought under tight statistical process control before being released to the plant floor. We do not mean to imply that there is no room for subsequent process improvement: the reverse is in fact true. But no plant should have to deal with process "debugging" during production runs. Second, production schedule freeze points should be set as far in advance as is feasible, working in close cooperation with the sales forecast and order departments. The goal is to reduce the level of fluctuations within daily production: the plant is a production process, not a buffer for sales forecast errors! We recommend a several-weeks-in-advance freeze point for at least the major volume and mix fluctuations, including both body styles and basic option levels. Many minor options, such as radio installations, can be changed a few days before production with minimal implications for production smoothness. In some cases, this discipline is enforced by sanctions: some plants actually pay fines to their upstream suppliers if they deviate from announced schedules by more than a given percent.