Lean Lives on the Floor
There are no gimmicks in lean manufacturing; to become lean, begin by mastering basic skills
By Art Smalley and Tom Harada
Lean Enterprise Institute
The Toyota Production System (aka lean manufacturing) has received lots of positive publicity over the past couple of decades in the US and other countries. The recent economic downturn and subsequent decline in company profits may dent some of that enthusiasm, but I suspect the attraction will resume and continue to grow in the future.
Despite many attempts at implementation, few companies have been able to match Toyota's success. This unsatisfactory situation is partly explained by the inherent difficulty of driving improvement. No one ever claimed implementing lean was easy. Also, however, we at LEI believe there is a lack of appreciation for the emphasis Toyota puts upon development and improvement of manufacturing processes in primary metal shops, such as casting, forging, machining, stamping, and welding.
Most early characterizations of TPS played up the material-flow aspect of the system. For example, as early as the mid-1970s, Toyota's prowess in manufacturing was being attributed to its famed "kanban" system. This view of things was severely limited, because in reality a kanban is nothing more than a simple tool used to control parts of Toyota's Just-In-Time system, and Just-In-Time is merely one component of the company's manufacturing system.
Later examinations of the TPS played up the system's kaizen aspect, and the importance of conducting improvement workshops. Some companies went so far as to establish quotas for the number of workshops that needed to be conducted to "become lean." Typically these workshops focused upon standardizing work practices, time and motion study, rearranging work flow, reducing the number of operators, shortening setup time, and attempting some form of one-piece flow. Many of these techniques have old roots in industrial engineering.
More recent characterizations of TPS have centered on the notion of a value stream, process flow, employee development and, of course, waste elimination. There is nothing wrong with these concepts or the techniques mentioned above. Unfortunately, as many practitioners are finding, on their own these concepts are often not sufficient to improve quality, cost, and delivery, especially in machine-intensive operations.
Our combined experience suggests that the TPS is much like the old 3-D mechanical Rubik's cube puzzle, which is not easy to sort out. Just looking at the cube from one angle will lead to frustration and inability to solve the puzzle. Portraying TPS as a material flow system, or even a system of human development, is necessary but not sufficient to consistently drive improvement. There are other dimensions that often must be emphasized. One of those dimensions is much more mechanical and machine-based in nature.
It's worth noting that TPS initially evolved in Toyota's engine plant in the 1950s and 1960s under the direction of Taiichi Ono, and not in the final assembly shops. Most textbooks use assembly-type examples (U-shaped cells, standardized work, parts organization) to depict TPS. Engine plants, however, depend upon machine tools, precision measuring equipment, jigs, fixtures, and tooling to remove material and make precision components such as crankshafts, cylinder blocks, or pistons.
Toyota's productivity level was estimated at 1/9th that of Ford Motor Co. when it began making improvements in the 1950s. Improvement required a long battle—initially fought in Toyota's machine shops—to catch up to North America in terms of quality, cost, and productivity. A trip to the Toyota Commemorative Museum of Science and Technology in Nagoya depicts the journey undertaken by Toyota since the inception of the company.
Nowhere in this museum will you find some of the more common and hyped tools of TPS that are touted in the US or other countries. Instead, this museum focuses on the importance of "making things," the "Spirit of Being Studious and Creative," and the importance of production technology. We believe the actual improvement journey that Toyota embarked on in its engine plants and other machine-intensive shops is grossly underappreciated, and that failure to understand what Toyota accomplished on the floor is one of the reasons companies often struggle to achieve gains in similar shops.
Some explanation may help illustrate this point. Occasionally we visit companies implementing TPS in machine-intensive shops, and are asked to give feedback about ongoing improvement projects. Usually the ongoing efforts focus on material flow in a value stream, scheduling, 5S, standardized work, or visual control. Sometimes a TPM or setup reduction workshop is underway. The problem is that machines continue to run with availability in the 50–80% range, quality varies widely, and scrap is common. Equipment-related delays are frequent as well. Operators are usually dismayed by the lack of connection between the problems they face on a daily basis and the ongoing improvement efforts driven from above.
We'll be blunt to make a point: Machine-intensive shops practicing these lean activities are more likely to struggle than succeed. Assume one company takes 10 processing steps to make a widget, and has 5% scrap, 70% uptime, rework, and other delays, and runs a fair amount of overtime. Another company has seven processing steps, near-100% availability, and 0.05% scrap and rework. Assume that labor, overhead, and material costs are roughly the same. Which one is in better shape? Which one would you rather manage?
The first precision machine tools in Toyota shops were imported from the US or Germany. Basically, all of these machines were operated on the one-man, one-machine basis that was normal for the time. Taiichi Ono, the company's manager of machining, embarked upon a strategy of breaking down this norm by having one operator run two machines, then three, and then four.
This methodology was highly effective, and helped Toyota close its productivity gap versus the US. Eventually, however, the effort ran into a wall. One person can only cover so many machines before you run into cycle-time barriers, quality problems, and minor stops between processes. In the 1960s, Toyota put great effort into purchasing and building transfer lines as a way to improve capital productivity. Some of the initial machines are still on display at Toyota's Commemorative Museum of Science and Technology. The plaques on this display, and others around it, proudly note the incremental advances Toyota made in building machine tools, jigs, fixtures, tooling, and measuring devices that drove capital productivity. This capital productivity can't be achieved merely by the superficial efforts we have outlined before—i.e., one-piece-flow, pull systems, kanban, 5S, and other lean techniques.
Consider a machining line that is no longer in existence at Toyota. It was replaced by a more modern and efficient layout about 10 years ago. In this layout for a crankshaft line, a single operator ran dozens of automated processes. More specifically, the line was all automated. The operator's job was to conduct periodic quality checks to audit the automatic ones, change cutting tools on a counter-based interval, conduct minor troubleshooting or preventive maintenance, and alert supervisors to problems requiring the help of either maintenance or engineering. The line ran at near 100% uptime and 100% quality.
How did Toyota arrive at this super-efficient machining system? Probably not the way you might think or might have read about in textbooks. This manufacturing line, and the one that replaced it, were based upon years of hard work in the areas of material removal, tooling, and machine and fixture design, as well as controls engineering. Toyota employs several hundred process experts worldwide who constantly gather data on the performance of current equipment, and work to build a better process in the future. Visitors to Toyota take the reliable and capable machines that they see for granted. However, such reliability and capability are the bedrock of the Toyota Production System. As an experiment, try running a pull system that involves standardized work with high levels of downtime or uneven quality. Unfortunately, it won't function very well.
Why does such process technology and capability go unnoticed by most observers of the Toyota Way or the Toyota Production System? We can offer up several educated guesses. For starters, consider the analogy of the iceberg and the ocean. The items above the surface in a Toyota factory are what you will notice during initial visits. Material flow, kanban, visual control, and standardized work, for example, are easy to spot.
Secondly, it's difficult to show the inner workings of a complicated production process. Toyota makes high-quality grinding machines at an affiliated company known as Toyoda Machine Works. These machines remove metal, and are controlled to the level of a few microns of dimensional accuracy. One of the keys to this abrasive-machining process is its patented hydro-stat main spindle bearing. The bearing is impossible to show, because it's inside the machine, and hard to explain to people not familiar with the basics of the process. So this type of feature deep inside the process goes unobserved by most visitors to Toyota.
Third, the details of the manufacturing process involve technical standards, as well as drawings and blueprints with specifications and tolerances. This type of detailed information is considered confidential by Toyota—and all other manufacturing companies. There is no incentive to show this aspect of TPS to the outside world. You can visit Toyota and get a copy of the standardized work chart with ease, but forget about obtaining any information that pertains to tooling, machine design, or fixtures. Toyota knows what is critical to their system and important to protect.
We suggest that persons looking to make improvements in machine-intensive shops not worry too much about the descriptions of TPS or lean manufacturing one sees in textbooks. Instead, consider your own problems and needs. The right move depends upon where you are in terms of capability and availability.
In general, we can make some comments and ask some questions that might help define a good starting point:
- In machine-intensive environments, make sure you can make the part right the first time. Jidoka (Buildin Quality) is not a pillar on par with the Just-In-Time part of the Toyota system by coincidence.
- The key to improving quality is often found in raw materials from suppliers. No pull systems and no amount of time spent standardizing the work routines of operators will solve supplier-quality problems.
- Internal in-process quality is usually affected by the quality of tools, jigs, fixtures, and critical parts of the machine. Toyota worries about maintaining 5 µm of runout (or less) in spindle heads on machine tools, for example. You probably have some critical factor in every process that deserves equal attention.
- Mechanical downtime must be analyzed and studied if it's to be reduced or eliminated. Root causes for downtime have to be pursued with the same level of rigor that is used in quality control, if you expect to improve.
- Efficiency (simply making parts per hour) should not be allowed to drive you to overproduce. Machines must be governed by the over-arching need to not overproduce.
- Taiichi Ono opined that success in TPS requires application of the scientific method. Learn to sort out "effects" from "causes" and problem-solve machinerelated issues. This skill will take you further than anything else that you will ever read about TPS.
Smalley teaches the workshop "Fundamentals of Lean Production—Creating Stability in the 4Ms" for the Lean Enterprise Institute, where he is a faculty member and author of the workbook Creating Level Pull: a lean production-system improvement guide for production control, operations, and engineering professionals, which received a 2005 Shingo Research Prize. Smalley also is president of Art of Lean Inc., and was one of the first foreign nationals to work for Toyota Motor Corp. in Japan in engine manufacturing.
This article was first published in the May 2009 edition of Manufacturing Engineering magazine.