Reducing variation in the products we makes should be an ongoing focus of all quality results. Reducing variation can be a long and arduous journey for many systems. Sources of variation in a system may be classified broadly into common causes or special causes. Special causes are typically thought of as those sources that cause a system to be statistically out of control. Common causes are those that are inherent to the system and when a system is in control we typically classify all the causes to be common.
How these causes actually get classified depends a lot on how you set up the sampling scheme to calculate the control limits. Also we usually think of elimination of special causes as easier to fix because we can identify the cause and fix it. To fix common cause you have to fix the system.
There is one source that can both special and common cause in nature. These are quality control strategies. For many systems business will define control strategies to compensate for shifts in the average of a quality response. These can be effective in reducing variation if designed correctly. Unfortunately most quality control strategies increase variation in the quality responses instead of reducing the variation.
There are straight forward ways of determining when a quality control strategy is a good idea and when it will be harmful. There are also ways to design the strategy that follow sound control loop type logic that will help insure success.
There are cases where we’ve seen reductions of 50% in quality variation when we’ve designed proper control strategies to replace the poorly designed control strategies.