9/1/1998 | 9 MINUTE READ

Operational Improvements: How to Build Cars More Effectively (Part 2)

Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

In Part I (August '98), Glenn Mercer of McKinsey & Co. describes some of the plant-level initiatives—such as team building—that are proven to boost productivity. Here he looks at the various departments of an assembly plant and provides recommendations regarding how operations can be improved.


Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

PART 2 of 2

Stamping or Press Shop Practices

The press shop is becoming a more critical part of the modern car plant, as body tolerances reach the 1- or 2-mm level in pursuit of ever-better panel fit, ever-stiffer chassis, and ever fewer squeaks and rattles. In fact, it is generally accepted that press shops are ideally located on-site with the rest of the factory, if feasible, to avoid the dimensional errors that creep in when stampings are transported long distances (including a surprising level of shipping-generated damage).


Press Shop Layout
Although much of the press shop's productivity is dictated long before the metal reaches the factory floor, there are some ways improvements can be achieved, such as through implementing extensive material handling automation and being vigilant in die maintenance.

Many of the drivers of press shop productivity are controlled by upstream design decisions, on both the product and process fronts. Therefore, the closest possible cooperation between car stylists, body-in-white engineers, and press process engineers is critical. Four measures indicate the likely level of success of the stamping operation, and reflect this need to integrate product and process design.

First, car designers must aim to minimize the number of stamped parts, and the number of press operations per part, to reduce the number of sources of eventual dimensions or surface errors. For a typical C-class car we would aim for less than 200 major pressed parts (half of them classified as medium to large in size), although each company counts these differently. The number of strokes per part must be kept to an absolute max. of 4, with an average closer to 2.5.

Next, there is the whole issue of automation, both its amount and performance. While a range of devices may be employed (e.g. automatic guided vehicles, cranes, robots), the material handling function should be automated as much as possible: for at least 95% of the steps involved. This is as much to improve quality as to boost productivity. Productivity is further improved by minimizing die change time (to about 3 to 5 minutes), and by maintaining a steady stroke rate (in the mid teens per minute) the reliability of the rate is as critical as the speed itself. As for the type of press best suited to the job, there is too much variation in car design and stamping philosophy to recommend a specific mix of transfer versus standard presses, especially when it comes to issue of asset flexibility.

Both these first two issues feed into the third key area, that of quality control. Product and process design largely drive press shop quality, but additional steps are of course taken in daily operations to maintain high quality. Die tolerances and their regular maintenance are certainly crucial, standardization of inputs vital (e.g. via minimization of the number of steel suppliers per product type), and process discipline vital. Generally one can target a level of rework below 1% as a gross measure of quality.


Body shop flow
Because the body shop is highly automated, it is essential that downtime be minimized.

Lastly, press shop inventory should be minimized in order to trap panel errors before they multiply. Two or three shifts' worth of material seems optimum in superior plants, although this amount depends on optimal batch sizing.

Welding or Body Shop Practices

Stamped steel sheets begin to form into a recognizable car shape in the welding shop, where hundreds of robots spot-weld the panels together. The sound practices here relate primarily to design, layout and configuration of the robots, as uptime and utilization of all this expensive equipment is the key to productivity.

A great contributor to productivity in the modern welding shop is product design, expressed in terms of how many main panels must be handled and how many welds must be executed (2,000 to 3,000 seems a good range, but this depends a lot on design philosophy, how welds are counted, and how many are included in supplied modules). The weld count itself is not so crucial from the point of view of time or materials, as a modern robot can add another weld quite quickly, but from the perspective of complexity of equipment set-up. Whatever car designers can do to ensure a manageable number and location of welds is helpful.

Since the typical body shop is very highly automated (at least 95% of the spot welds should be done automatically, and virtually all the material handling, to avoid panel damage), equipment uptime becomes a critical factor. We find total downtime of less than 3% to be sound practice, with additional line losses kept well below 10%.

Manual rework of erroneous parts should also be minimal if the welding process is under control with less than 5% of total time in the shop. Scrap, on the other hand, should be virtually non-existent, given good automation and programmed process control.

Paint Shop Practices

The paint shop is in some ways the most mysterious of the major operating areas of a modern car plant, as it is always sealed off to the outside world in order to keep any dust or dirt from the freshly painted car bodies. Also, expensive and extensive robotic and computer-controlled equipment rules here, making it difficult to assess productivity and quality. Then there are different philosophies among companies about which types of paint to use (e.g. solvent- vs. water-based) and how many layers to apply, etc. But a few good practices seem to be clear.


Paint Shop Typical Layout
Quality requirements result in the need for high levels of automation in paint - 90% or more. Given the demands of the consumer with regard to the painted surface, expect a first-time-pass failure rate of up to 10%.

First, high levels of automation are required here, both to spare workers exposure to noxious fumes, but also to ensure quality control. An automation level of 90% or more is necessary in paint application.

Rework is almost inevitable with a high-customer-impact, fickle material like paint. An up to 10% first-time-pass failure can be tolerated. Rework can be minimized at the design end (by ensuring "congruent" panel fit) and also by continuous problem-solving in the shop itself (e.g. correcting paint dripping).

The nature of the paint process cycle is small-batch as opposed to stamping's large-batch and welding's and assembly's one-by-one process.

Assembly or Trim Shop Practices

It is here, where the car's parts are assembled into the painted body, that most of the complexity in the vehicle emerges, and where the greatest amount of labor is employed. Automation recedes into the background (few trim shops are more than 5% automated, by job count), except to relieve workers of excessive risk or strain. Thus, human factors dominate the productivity equation. Critical good practices are numerous.

There must be broad implementation of design for assembly (DFA): product development must do its job for the trim shop by keeping separate part counts as low as possible, moving to good DFA practices whenever possible (e.g. snap fits), and minimizing awkward, hidden, or overhead assembly configurations. Factory workers everywhere can attest to the value of good DFA: for example, the first-generations Ford Taurus was famous for being so easy to assemble that it "fell into place."

Work design and organization are vastly important: books and books have been written on this topic, many discussing the nuances of the Toyota Production System. The key goals to point out here are:

  • zero line balance loss
  • minimization of waste of all kind (scrap, space, motion, etc.)
  • simplified and standardized work procedures
  • high worker involvement in task design
  • steady, continuous daily improvement in the process.

Rework must not only be minimized but not tolerated, and should be in the 3% range at most (at line end). If rework rises above this level, it implies processes are out of control and must be rectified immediately.

As noted, automation is not critical in final assembly, but we found the better plants using quite a bit of low-cost, low-tech automation or other powered equipment to offset dangerous or stressful work (e.g., tire/wheel mounting) or to ensure critical quality targets are met (as in fitting and sealing the windscreen to the body).

Modular assembly is becoming quite a trend in many plants around the world, reaching a peak in the Mercedes Smart-car plant in France and in the GM "Blue Macaw" factory in Brazil. While there is debate as to how far to take modular supply—wherein more of the car is delivered in pre-assembled chunks—it is clear that the advantages still can outweigh disadvantages. These advantages include the smaller amounts of precious labor, floor space, and inventory tied up in the car plant, lower wage costs at the supplier's staging facility, and the ability of the supplier to redesign all module parts for superior integrated performance. Barriers to modular assembly include the complexity of design and concerns about concentrating so much value in a smaller number of supplier firms, but the trend toward modularity indicates these issues are being dealt with. Typical assembly modules seen today include seats, "stuffed" instrument panels, and fuel tank assemblies.

There is much more to be said about the myriad details that go into a smoothly functioning final assembly operation, so these five points should be considered samples rather than an exhaustive list.

Service and Support Functions Practices

These functions will vary significantly by plant (e.g., according to whether there is local engineering support). A few principles apply that separate the better from the average performers at the plant level. We are not in this discussion covering headquarters functions, which are a separate topic entirely.


An Assembly Layout
Final assembly and trim are the least-automated areas in an automotive plan. Workers can be aided not only by devices in the plant, but also by the implementation of design for assembly (DFA) methods prior to the product hitting the factory floor.

First, outsource where you can, to concentrate on the core business of making cars. Cafeteria, laundry, general maintenance, security, medical service, basic data processing, payroll, etc., can all be handed over to outside firms for efficient and effective execution, to relieve plant personnel of worrying about them.

Next, focus your information technology (IT) resources on the integration of the production process. It is less important that plant IT be cutting edge (although we strongly recommend client-server architecture) than that it work to integrate the thousands of parts, thousands of workers, and hundreds of suppliers the plant must deal with daily. At a minimum, suppliers must be fully electronic data interchange (EDI) integrated, and the plant part system must be transparent to the process controllers (as opposed to the sometimes mysterious output of older MRP II systems).

Finally, although it seems an obvious point, plant administration must be stripped down to only those services that clearly add value to the assembly operation. Accounting must be on site of course, but it should be working to develop a management-oriented activity-based costing system, as opposed to just generating reports for senior management. Personnel or human resources must be there to handle legal requirements and do some training, but team member selection and the definition of member qualification standards might be better delegated to the teams themselves. Any functions not specifically supporting the plant should be moved offsite and away from plant management, to avoid any distraction.

We hope these remarks have proven useful to you as you consider your own plant's situation. These comments are top-level summaries. There are many discrete details to take into account, as well. Also, please be aware that the car business is an evolving industry, and so no "best practice" stays "best" for long!


  • The Changing Definition of 'Niche Vehicles'

    Once the playground of exotic car makers, the definition of a niche vehicle has expanded to include image vehicles for mainstream OEMs, and specialist models produced on high-volume platforms.

  • Designing the 2019 Ram 1500

    Ram Truck chief exterior designer Joe Dehner talks about how they’ve developed the all-new pickup. “We’ve been building trucks for over 100 years,” he says. “Best I could come up with is that this is our 15th-generation truck.”  

  • Cars Then & Now & in Cuba

    This is the very first Chevrolet Camaro: It has VIN #100001.