The dictionary definition of quality is “the nature of excellence in workmanship.” Kind of vague, yes? Over time it has been reinterpreted to “conformance to specifications,” with its most recent mutation being “parts to print.” It is important to understand the history of this definition.

The master craftsman was an individual competent in a thousand tasks, meaning things were built one at a time. Tolerances didn’t matter because the master made things work by custom-fitting the pieces together. But the industrial revolution gave us a thousand people competent at one task—the beginnings of mass production.

Mass production no longer had the luxury of fitting/adjusting things together that didn’t quite go together. While most people give Henry Ford credit for mass production of standardized parts, it actually started with a defense supplier. Eli Whitney, while given credit for inventing the cotton gin, also owned the Whitney Armory, a manufacturer of rifles. One of the many contributing factors to winning the Civil War was that if you had a Whitney rifle and it broke, you needed only to find another one and take parts from it. Whitney’s interchangeable parts could only work if they were all the same. This gave rise to the first toleranced drawings in the 1850s.

In the following years, as machine tools became more precise, we had standard drawing conventions and tolerancing. Frederick Taylor and Frank and Lillian Gilbreth introduced time and motion study and standardized work methods, further contributing to part standardization around the turn of the 20th century. Henry Ford put mass production to work during WWII, further refining tolerancing and standardization.

Keep it simple, Mikhail
Here is where history conveniently misses an important quality philosophy. Most historians of lean look to W. Edwards Deming and Joseph Juran as the founding fathers of statistical quality during the reconstruction of Japan after WWII. Their work spawned the philosophies and massive publications of Ishikawa, Toyoda, Ohno, and Shingo that ultimately evolved into the Toyota Production System and what we now call lean.

What’s missing? During WWII, Russia was left to itself to provide arms for its massive army. What was needed was a gun a peasant soldier could use with minimal training, made inexpensively, that would operate in the rigors of the Russian weather. The German rifles were precise, accurate, and deadly. However, they didn’t work too well when dirty, frozen, or wet.

While in the hospital, a wounded tank mechanic modified a German-designed submachine gun and improved it. Instead of using exotic materials that he had no access to, or tighter tolerances that the Soviet machine tool industry couldn’t reproduce, he used basic materials, simple manufacturing methods (mostly sheet metal stamping), and introduced the concept of a robust product with wider, less precise tolerancing. Mikhail Kalashnikov’s iconic AK-47 basic design hasn’t changed in six decades. Because of the loose tolerances, it is more reliable than any other automatic rifle today.

If you read the great writers on quality, they all promote the widest possible tolerances to generate the most reliable product. However, readers of Lean for Dummies might think higher precision (smaller tolerances) means a better part.

Making a part that works
Before we get to SPC, TQM, 12-Sigmas, and all the other silliness, you have to ask yourself, “What is an acceptable part?”

The definition of an acceptable part breaks down to three components. To keep it simple, it’s what your significant other thinks of you:

1. You do what you’re supposed to do when you’re required to do it (functionality).
2. You are sufficiently good looking for the purpose you serve (cosmetic acceptability).
3. Size doesn’t really matter so long as #1 and #2 are fulfilled (dimensional consistency).

Most of the hoopla in quality focuses on size (#3) but ignores functionality (#1) and cosmetics (#2). The reason for this is simple. We all can measure size, but agreeing on functionality and cosmetics seems to be a gray area in the minds of the quality lean ninjas.

Cosmetics (#2) are simple: The SPI has published Cosmetic Specifications of Injection Molded Parts. Cosmetics is not a beauty contest. There are very rigid standards. All you need to do is pick the one you want your part to comply with and stick with it.
Functionality (#1) is also simple, but designers hate it. They must devise a binary test where the product either passes or fails under end-user conditions. There is no such thing as conditionally passing or marginally failing. It either works or it doesn’t during its intended design life. Period.

The other thing designers hate is that if it works, it must have met the design intent, regardless of the dimensions. This usually means massive engineering changes to make the dimensional specifications line up with the part’s variances and dimensions. This is the “D” part of R&D; but in the rush to market, it is the most easily ignored and hence the obsession with the idea that size matters.

Nobody (and nothing) is perfect
When I am challenged by the 12-Sigma lean black belts for the causes of waste, I always ask two questions:

#1: What’s a good part? They readily reply wrongly with the technogarble of Cpks, geometric tolerancing, SPC calculations taken from a digital layout machine, etc., etc.
#2 (the ship sinker question): How do you know if it isn’t precisely “to print” and/or not absolutely “free of manufacturing defects” (Cpks, geometric tolerancing, SPC calculations taken from a digital layout machine, etc.) such that it will fail 100% of the time in use, resulting in massive lawsuits and monumental product recalls? This question is met with silence, and is when I’m generally labeled as “not a team player.”

A good part is something that absolutely passes functionality and is cosmetically acceptable to the end user. A bad part is something that either gives the impression of such low-quality workmanship that the customer won’t buy it in the first place, or doesn’t meet the end user’s expectation of what was purchased, resulting in them demanding a refund.

If you study the Toyota way in practice, you’ll see Toyota is very focused on functionality and acceptable (not perfect) cosmetics. While probably under the guise of Deming, Shingo, or Ohno, they are actually following the design philosophy of Kalashnikov—make the design both robust and dimensionally forgiving. This will:

• improve quality
• reduce the need for excess inventory
• reduce the need for rework
• eliminate scrap
• reduce overproduction
• shorten the time to accomplish the task
• reduce total costs.

These are the big time goals of lean, and all are achieved simply by getting quality out of the beauty contest business and ensuring that manufacturing makes something that works. If you can’t revamp your quality procedures (notice I didn’t say upgrade, tell you to slap another Sigma on your ninja’s black belt, or ask you to make another chart), just keep doing what you’ve been doing and save your money by not doing lean. But if you can do the revamp, start thinking about all the extra profit you’ll be making.

How to specify a “good part”
After everyone has agreed on the criteria, classify the surfaces of the parts according to a cosmetic specification—highly visible, rarely visible, and so on—with a time and distance specification that considers the eyes of an end user as to whether or not it is acceptable. Assume the part is cosmetically acceptable. Any inspection should be done with the mentality that whatever changed has crossed the threshold of acceptability.

Take the dimensional specification of a functionally approved part and pick a few easily measurable dimensions. The assumption here is that if it passed functionality with the range of dimensions you provided when the samples were submitted, the only dimensional criterion will be that, if the dimensions have changed, the process has changed, and therefore functionality may have changed.

Quality must rethink its mentality and begin with the assumption that all parts produced are acceptable. Its purpose in inspection is to maintain process consistency. Think this way and being lean results in the profits rolling in.

This is the second of three articles on lean (read the first here). Installment #3 will discuss lean and JIT. Consultant Bill Tobin is a regular contributor to IMM. You can sign up for his e-newsletter at www.wjtassociates.com.