The operations manager and I were touring the final assembly line at an engine plant in the Midwest. We stopped at a workstation, one of many, and he pointed out to me a torque wrench hanging there. He explained that this workstation had been a source of quality defects related to improper torqueing in the past. The solution to this issue had been to install a networked tool that not only would torque correctly and consistently, but would also be connected to the factory MES system, store the torque data for that specific serial number, and not permit the line to advance if there was any problem.

I asked if that was a solution to the quality problem. Yes, he responded, but there was a significant capital investment, there is now an additional maintenance requirement, and another potential source of line stoppage if the system malfunctions. So this fix was something of a mixed blessing, but it was a common response to quality problems: look for a technical solution that takes the responsibility away from a human being.

A few hours later and only a few miles away, another factory faced similar challenges. At this factory the initial instinct, once a quality issue has been identified and contained, was not to attempt to error-proof the work step with technology, but rather to look to other options first. Are changes needed in the Standard Work Definitions? Was there lack of training for the operator? Could a simple fixture or tool provide a fix? Could operators provide insight into the issue, and help with a solution? Technology and automation may be used, but not as a first choice.

This article will present the Lean approach to quality, especially within the domain of human beings doing work. Included will be a presentation of both the Lean philosophy related to quality, and some of the specific tools and methods currently in use by Lean leaders. Machine-related quality, while critically important, is outside of the domain of human beings doing work for purposes of this article.

Let’s walk through the process of responding to a quality-related issue in a Lean factory, as I have experienced it. We can then explore the various quality-related tools and methods that a company may choose to use. When a quality issue is discovered, either on the production line or in a final inspection process, the first action is to contain the problem, and keep it from spreading or affecting other products. This requires an immediate response by team leaders, supervisors, and engineers, and it may also require implementing processes or procedures that are not intended to be permanent. Until the root cause of the issue is determined and addressed, for example, a temporary inspection or test may be added to make sure that the problem does not continue. Once the problem is contained and not allowed to spread, the next step will be to identify the “root cause.” This can often be accomplished with a simple investigation; in other cases data will need to be collected and analyzed, and experiments conducted. Daily meetings with supervisors and team leaders are focused largely on these efforts. The main point here is that until the “real” or root cause is found and addressed, the problem is likely to reoccur.

In a Lean culture, the actual fix of a problem is different from that in a more traditional manufacturing culture. In the engine plant mentioned previously, there was an inherent lack of trust in the workers, with a strong bias towards removing control from the human, and applying a technical solution. In a Lean culture a technical solution may end up being the best one, but the first instinct is to rely on the operators (remember that we’re discussing operator workmanship here) as the first line of defense, and to look for simple human-centered solutions first.

There is also a striking difference in how problems are thought of in a Lean culture. Toyota refers to problems as “jewels”: precious opportunities to improve. If there are no problems, there is nothing to work on for improvement, and in this sense problems are to be treasured. In fact, Toyota managers begin to get concerned if the number of reported issues (small and large) drops too far. They see this as an indication of two possibilities: that operators are encountering problems but not reporting them, or that the production environment is not challenged enough and there is too much fat in the system. There is a sweet spot in the number of reported problems, where production is sufficiently challenged and improvements are constantly being made.

Here are some of the tools and methods that a Lean culture will rely on, starting with the most important one.

Standard Work

This is the foundation for everything that follows when discussing the Lean approach to quality. The Standard Work Definition describes the work steps within a process, the expected work content times, and the material consumed. In the realm of quality it also includes two important data elements: Quality Criteria and TQC Quality Check. The Quality Criteria answers the question “Could something go wrong at this detailed work step”? If the answer is yes, then some additional instructions will be needed regarding how to do the step correctly. The Quality Criteria breaks the work content into two pieces: the work itself, and the quality related to the work. It may include a quick Self-Check, performed by the operator, to ensure that the work was done correctly before moving on to the next step. The TQC Quality Check is an indicator of the need for a second check of this same step, a second set of eyes, to ensure that a defect does not get beyond the station where it occurred.

Training

Operator training is the next logical step once Standard Work has been initially defined. Training takes place under the mentorship of a team leader or supervisor, with an eye on achieving three things: a sufficient level of skill in doing the work, in being able to do quality work within the expected standard time, and in being able to check one’s own work for quality. Insisting on Standard Work, far from taking responsibility away from the operator, is the most respectful thing you can do for a new hire. It allows them to apply the best ways of doing the work that are known to the company, safely and ergonomically, and creates the foundation for being able to make meaningful improvements to the work once the standard has been mastered.

Visual Work Instructions (VWIs)

Sometimes also called Graphical Work Instructions, this tool is the conversion of text-based instructions into operator-friendly illustrations. The goal of VWIs is not to replace detailed text-based documentation, but to provide easy-to-use quality-oriented reminders to already trained operators. In high-mix production lines where a particular unit may not have been produced in a while, a visual reminder of the work step and quality criteria is very helpful, without having to read through detailed text-based Standard Work Definitions.

It is possible to go overboard on the use of Visual Work Instructions, especially if the work is repetitive and simple. An American faucet-maker went to considerable time and expense to create VWIs for their many different products, where the work content itself was small and very repetitive. Operators didn’t need them, and ignored them. Apply common sense to their creation and use.

Check-Do-Check

Even with quality checks identified and workers trained, humans will make mistakes. At a minimum, if the Standard Work Definition calls out a Self-Check because an error is possible, the operator doing the work will check his/her own work. If the frequency of occurrence or risk is relatively low, this may be sufficient. If the frequency or risk is higher, an additional TQC Quality Check will be needed. Here’s how it works:

If a work step is quality-critical, and the risk of a defect occurring is medium or high, then it will not be sufficient to have a single operator check their own work. As we know, the chance of a person finding their own mistake is less than if a fresh set of eyes performs the check. The TQC Quality Check is an indication that in addition to an operator-performed inspection, it will be necessary for a second person to double-check that same work step. This would normally occur at the next workstation, so as to not disrupt the flow of work. If the step cannot be checked at the next station, then a Call-Over TQC may be needed to do the check sooner. This could be done by a Team Leader, or by calling over a neighboring operator. This is obviously less desirable, since it is disrupting to normal work, and may be practically impossible in a high volume line.

Is this actually effective in catching and fixing errors? If an operator makes one mistake every 100 opportunities, and does not check his/her work, then the defect rate will be 1%. By formalizing the Self-Check step, i.e. actually checking his/her work, the operator can still fail to catch defects, but at the same level of quality (missing one out of 100 checks), the overall defect rate drops to 1/100 * 1/100 or 1/10,000. Adding a second set of eyes, again at the same level of quality (missing 1 out of 100 checks) will further reduce the defect rate to 1/100 * 1/100 * 1/00 or 1/1,000,000. Don’t take these numbers too literally, but there is no doubt that checking work, and especially including a double-check, will have a powerful effect on catching errors.

Touch For Quality

The Touch for Quality method is an enhancement to the Check-Do-Check method that can also contribute to a huge drop in workmanship errors. As opposed to simply looking at an inspection point, operators are required to physically touch the inspection point, either with their finger or with a pointing device. There are two advantages to touching instead of simply looking. First, it helps to focus the attention of the operator since it involves a physical action. The quality of the check will be better. Second, it allows Team Leaders and Supervisors to confirm visually that the check is actually being done.

The pointing device could be something as simple as a wooden chopstick. A speaker factory in California was having trouble with a gasket installation, and introduced the Touch For Quality method using chopsticks from their local Chinese restaurant, to both help with the gasket installation as a tool, and then to inspect it to make sure it was installed corrected. Yield at the end of the line improved by 300%! An Indonesian factory experienced such an improvement in workmanship quality after implementing Touch For Quality that they began to call the pointing devices “magic sticks.”

Error-Proofing

The difference between Lean error-proofing and traditional error-proofing relates to the Taiichi Ohno saying: “Use your brains and not your money.” This is a big topic, but in essence it means finding simple ways to eliminate or greatly reduce the possibility of errors. The Japanese term for this practice is poka-yoke. The library of books and examples of error-proofing is extensive, so I’ll simply mention it here.

One of my favorite examples of a poka-yoke tool involved a simple modification to a screwdriver. A screw needed to be installed to a specific depth in a hole, and the effort to tighten, then measure, then loosen was very inefficient. By cutting a slot in the side of the screwdriver blade to the correct depth, the screwdriver could not only do the work, but also function as an inspection tool, without requiring the operator to put it down.

Failure Mode and Effect Analysis (FMEA)

FMEA is a well-known methodology, and takes analysis of Quality Criteria to a deeper level. Every work step is systematically reviewed to identify the potential Failure Modes (what could go wrong) and the potential Effects (what would happen if that defect occurs). In addition it is necessary to know if the defect is detectable. If the defect occurs but it is not detectable, that is much more concerning than if the defect is obvious.

The beauty of the FMEA method is that it allows you to create a score, based on the relative importance of Frequency, Effect, and Detectability. Not all potential defects are equal, and FMEA allows you to identify and work on the biggest offenders first.

Operator-Based Kaizen

The involvement of operators in improving processes, by contributing small improvement suggestions, is an aspect of Lean that is often misunderstood. While it is true that Lean organizations are improving quality and eliminating waste through these efforts, the primary purpose is not cost-reduction or error-proofing. Instead the suggestion program’s main objective is employee development, and training the workforce in problem-solving. The Lean slogan heard at Toyota is “monozukuri wa hitozukuri,” or “making things is making people.” Operators, the people actually building the products, are on the front-line for quality, and operator-based Kaizen is an operator-development tool.

Simply announcing that operators will be expected to suggest improvement ideas doesn’t make that happen. Toyota Material Handling in Columbus, IN, admitted that it took years to develop their internal culture. When the plant opened in 1990, the idea of operator involvement was a new one for the workforce, and met with some distrust. Only by finding improvement champions, and vigorously supporting their initiatives, was this plant able to overcome skepticism over time, and achieve their current average of between one and two suggestions per person per month.

 In this article we have reviewed some of the main quality-related tools and methods in use at Lean factories. Technical or capital-intensive systems and methods can make an important contribution, but we have seen that the initial instinct in a Lean factory is not to take responsibility for quality away from the people who do the work, but instead to look initially for simple and inexpensive improvements. This may be another reason why Lean companies also tend to be the most profitable in their industries.