OR WAIT null SECS
Where is the variability coming from and what have we done to minimize it?
Textbooks and journal articles treat common cause variation as if it is an inevitable fact of nature and beyond our control: "In any production process, regardless of how well-designed or carefully maintained it is, a certain amount of inherent or natural variability will always exist. This natural variability or 'background noise' is the cumulative effect of many small, essentially unavoidable causes" [Emphasis added] (1). This attitude cuts off thoughts of trying to reduce variation. But, with some reflection, there are several ideas and techniques that can begin to help reduce common cause variation.
Work to hit the target
Aiming for and hitting the target, whether it is x, y or z, seems a simple idea, but it could be argued that it is everybody's responsibility to know what the target is and to do everything possible to hit that target every time. One person achieving the target infrequently doesn't help. But 400 people hitting targets a dozen times a day can have a dramatic effect on reducing variability. The target could be something as simple as setting the temperature on a dryer or as complex as a management objective. Keep in mind that a specification range is not a playground for manufacturing.
Lynn D. Torbeck
This mindset also can influence how specifications criteria are written. It is common to write the criteria as "(Low, High)." I suggest that criteria are better written as, "Target (Low, High)." The first thing the operator sees is the target value; people generally try to achieve the first thing they see. Second, the low and high criteria limits are given, which eliminates the need to mentally calculate those limits (e.g., if Target ±Δ, was used instead). The limits need not be symmetrical with the target value.
The terms sound contradictory but do contain logic. Many activities are at the liberty of the operator or analyst and, as such, are subject to considerable leeway in how they are performed. In these situations, particularly in the analytical laboratory, one strives to get everybody on the team to do exactly the same thing, the same way, every time. If then, at some point in the future someone proposes a new or better way to perform the task, the whole team changes to follow the new process. This group consistency can have a substantial impact on variation within a department. Notably, creativity of the individual is not stifled, but rather channeled to find better ways to perform a task.
"An operational definition describes what something is and how it is measured" (2). For example, "sample the batch" could mean: "Using the 72 in. thief, open the port on the right side labeled P8, and take a sample from the top two inches, a sample from the middle, and a sample two inches from the bottom. Composite the three samples into a clean glass jar with a lid, and label with the date, time, name, product, lot and vat number." Operational definitions reduce variation by promoting consistency. Standard operating procedures are a form of operational definitions.
Mistake proofing or poka-yoke
Made famous by the Japanese auto makers, poka-yoke is simple but powerful in reducing variation, deviations, and discrepancies. The goal is to make activities as mistake-proof as possible by physical means or by procedures that are difficult to do incorrectly. The classical physical example is to put a mechanical stop on a drill press to prevent the drillbit from making a hole that is too deep. In a paperwork process, colored pages are used to clearly identify certain documents.
Control what can be controlled
Although controlling what can be controlled may appear to be an obvious idea, many factors are commonly ignored during normal operations. Perceived to be noncritical process parameters, they are left to float within some specified range. However, variables should be controlled to the fullest extent with the highest accuracy possible without incurring great expense or requiring great effort. Again, controlling only one factor will have a trivial impact, but a culture of controlling hundreds will reduce common cause variation. The tools discussed in this article require support from management, but it is that support that makes implementation so powerful.
1. D.C. Montgomery and G. C. Runger, Applied Statistics and Probability for Engineers (Wiley, New York, NY, 1994), p. 834.
2. P. R. Scholtes, The Team Handbook (Joiner Associates, Madison WI, 1988), p. 2–28.