Many quality teams stop at one sentence: 80 percent of problems come from 20 percent of causes. As a rough illustration, that sentence is useful. As a rule, it is misleading. The numbers 80 and 20 should never be a threshold to chase, and the person who named this principle actually stated that the name was incorrect.
The name Pareto was assigned by Juran, then corrected by himself
Vilfredo Pareto (1848 to 1923) was an Italian economist who studied the inequality of wealth distribution and developed a mathematical model to describe it. The observation that elevated this to a general principle was made by Joseph M. Juran in the early 1950s when he prepared the manuscript for the first edition of the Quality Control Handbook. Juran needed a short name and used Pareto's name.
In 1974, Juran published a paper titled The Non-Pareto Principle; Mea Culpa and acknowledged that he had assigned the wrong name. Pareto's work was in the realm of economics, and his model was not intended to be used in other fields. The cumulative curve that is still attached to the Pareto diagram was not created by Pareto but rather a form developed by M. O. Lorenz between 1904 and 1905.
Juran also changed his own terminology from vital few and trivial many to vital few and useful many, so that the small remaining category would not be considered worthless.
The practical consequence is one sentence: do not force your data to land on the numbers 80 and 20. What matters is the distribution's skewness, not how skewed it is.
What matters is the cumulative column, not the number 80
A usable Pareto diagram has three elements: contributors sorted by their largest contributions, the amount of each contributor's contribution in numbers, and the cumulative percentage of the total.
The third element is the one that is most often missing, yet that is where decisions are made. The percentage of each category is calculated by dividing the number of occurrences of that category by the total occurrences, then multiplying by 100. The cumulative percentage is the sequential sum from top to bottom.
A documented example from Juran's archives: a team examined an order form containing 18 items and recorded the frequency of errors over one week. The top four items contributed 29, 25, 21, and 11 percent of the errors. When summed sequentially, they accounted for 86 percent of all errors. Four out of 18, not exactly 20 percent, and not exactly 80 percent. The number that makes the decision reasonable is the 86, not the 80/20 slogan.
If there is no clear break
Sometimes the bars are nearly the same height, and the cumulative line rises gently without a break point. Juran referred to this condition as the Awkward Zone. This is not a sign that the principle does not apply, but rather an indication that the basis for grouping is not yet accurate.
What needs to be done is to change the way the data is classified and then rearrange the diagram. Some grouping bases that usually yield results include: type of error, time of occurrence by hour, day, or month, order of process steps, type of product or service, location of work, and characteristics of the group performing it.
Two traps that often go unnoticed
First, the bar labeled "others." Combining the remaining categories does tidy up the appearance, but it hides the reality that each of those categories contributes too little to be addressed individually. If that bar is sorted as if it were a category, the order becomes incorrect.
Second, counting the number of occurrences while what is at stake is cost. A cheap defect that occurs 300 times and an expensive defect that occurs 30 times will place the cheap one at the top, even though the repair that saves money is next to it. The Pareto diagram can be constructed based on cost values, not just frequency. Juran illustrated this with a car transmission factory that constructed a Pareto from Cost of Poor Quality data to select which cost categories to address first.
Pareto points the way, not the cause
The Pareto diagram is one of the seven basic quality control tools introduced by Kaoru Ishikawa in 1968, along with check sheets, cause-and-effect diagrams, histograms, control charts, scatter diagrams, and stratification. It tells you where to look, not why the problem occurs, and it is only as good as the data that fills it. Therefore, it is commonly paired with cause-and-effect diagrams or root cause analysis in the next steps.
The 80/20 principle in production lines is also discussed by Muhammad Husni Mubarok in the QA/QC Operation class. If you prefer to see his explanation directly, there is a clip on the video page.
Sources
- The Non-Pareto Principle; Mea Culpa, J. M. Juran (1974), Juran Institute archives
- Pareto Principle (80/20 Rule) and Pareto Analysis Guide, Juran Institute
- Understanding the 7 QC Tools: Main Tools in Quality Control, PQM Consultants
- QC Seven Tools (Seven Quality Control Tools), IPQI
- Pareto Diagram: Understanding, Function, How to Create, Kwik Kian Gie Campus