Zero Defects – A Realistic Goal or a Dream?

Achieving zero defects is not a dream at all, but rather a realistic goal to be pursued.  We expect and demand zero defects all the time in our personal life.  For example, how often is it acceptable for a plane to crash instead of successfully landing at the destination airport?  How often is it acceptable to eat a meal at a restaurant and get food poisoning?  How often is it acceptable for our bank account to lose some of our money?  How often is it acceptable for someone else’s charge to show up on our credit card account?  In all cases we expect that these defects will never occur.  We should have the same tolerance for defects at work that we do in our personal life – zero.
In order to move ever closer to the goal of zero defects, we must understand how defects occur and how to prevent them.
Early in my career I was an Inspection Supervisor working in a forging plant in the automobile industry.   The production department was tasked with making parts.  The inspection department was tasked with visually sorting out the good parts from the bad ones, and then applying sampling inspection to make a decision about whether or not to ship a lot of (supposedly) good parts based on an acceptable quality level (AQL).
Studies have shown that visual inspection is roughly 70% effective, meaning that the batches of parts that were deemed ready to ship would almost certainly still contain defects.  The acceptance sampling plan that we used, based on Military Standard 105D, called for a sample to be taken from the lot and evaluated for defects.  The size of the sample was based on the size of the lot.  The number of defects that could be present in the sample while still considering the lot to be acceptable was defined based on the AQL.  In other words, a decision could be made to ship the lot to the customer even though it was known to contain some level of defective parts.  This approach seems even more absurd in retrospect because many of the parts were critical to vehicle safety and were regulated by the Motor Vehicle Safety Standards legislation!  (Note:  Mil-Std-105 was officially cancelled in 1995).
Judgment inspection (making a product and then evaluating whether or not it is acceptable by comparing it to a standard or requirement) is a costly and ineffective approach that can only hope to find some percentage of defects after they have already occurred.
In Six Sigma we talk about the output of a process (the Y variable) being a function of the inputs to the process (the X variables).   A defect in the Y output of a process is a function of an error in one of the X input variables that impacts the Y output.
A much more powerful approach for defect reduction and elimination is source inspection.  Source inspection is used to detect the errors that lead to defects.  If an error that could lead to a defect is detected at the point where it occurs (the source), the process is either automatically stopped for correction or automatically adjusted so that the error condition is corrected.  Defects are reduced or eliminated because the errors are never allowed to turn into defects.
For example, think about baking a cake.  The Y output that is desired is a cake that is fully cooked without being burned.  Two important X variables are the temperature and the length of time that are used to bake the cake.  Source inspection would be applied to this process by automatically controlling the temperature and the baking time to the proper levels.
Here is an example from my consulting experience.  I was working with a company that prints cereal boxes and other food packages in huge quantities.  A common defect was that the printed image was blurred.  Judgment inspection was being employed on a routine basis to attempt to sort out the good boxes from the bad ones after they had been printed.  Statistical process control was being applied, but only to the output of the process.
When I asked about the error condition that resulted in the defect, I was told that it was the amount of tension on the roll of paper that was feeding the printing press.  If the tension was too high or too low, the blurred image would result.  I then asked what was being done to monitor the tension and make sure that it was correct.  I was met with blank looks and was told “We don’t monitor the tension”.  The folks that were running the process knew when they were making bad output, but they were not monitoring the error condition that resulted in the defect!
The proper approach in this case was to install a tension monitoring and control device on the printing press that automatically kept the proper amount of tension on the paper, thus eliminating the cause of the defect.
There are two additional approaches to eliminating defects that should also be considered.  One is to combine source inspection with defect detection.  In the example of the printing of cereal boxes, we combined tension monitoring and control (to eliminate the error that led to the defect) with automatic vision detection of the defect (blurred images).  If the defect was detected, the process was automatically stopped for correction so that defects would not continue to be produced.
The most definitive way to prevent the defect from ever occurring is to eliminate the source of the error by design.  An example that I often use in the classroom is the design of the common household plug on the end of an electrical cord.  A two-pronged plug is designed with one prong wider than the other so that it can only be plugged in with the correct polarity.  Another example is the design of a railroad crossing.  In the case of a railroad crossing that is guarded using signal lights and barriers, the error condition is that a vehicle ignores the barriers and lights when a train is approaching.  The defect is that the vehicle is struck by the train.  If the design is changed such that the vehicles cross under or over the track, the chance that the defect can ever occur is eliminated.
Your comments or questions about this article are welcome, as are suggestions for future articles.  Feel free to contact me by email at
About the author:  Mr. Roger C. Ellis is an industrial engineer by training and profession.  He is a Six Sigma Master Black Belt with over 45 years of business experience in a wide range of fields.  Mr. Ellis develops and instructs Six Sigma professional certification courses for Key Performance LLC.   For a more detailed biography, please refer to