Lean
Who Invented the Pareto Chart?
Vilfredo Pareto, Joseph Juran, and Kaoru Ishikawa are all contenders
Published: Monday, June 3, 2019 - 12:03
When I first learned quality improvement back in 1989 at Florida Power and Light, the consultants who trained us taught a very specific way to draw a Pareto chart. They’d been trained in Japan, the place where quality improvement first took root during the 1950s, so I took it for granted that the way they drew Pareto charts was the authentic and best way to do so.
A Pareto chart combines a bar graph with a cumulative line graph. Using the way we were taught to draw a Pareto chart (figure 1), the bars are touching, making it extremely easy to visually compare levels from one bar to the next. The bars span the entire available space along the x axis. The cumulative line graph springs from the bottom left corner of the first big bar, and each subsequent point is plotted from the corresponding top right corner of its bar.
 Figure 1: Recommended Pareto chart |
As I continued learning about quality improvement, I occasionally came across Pareto charts drawn in a different, far more confusing, manner (figure 2). The bars stand far apart from each other like a picket fence, making it difficult to compare the levels displayed by the bars. The cumulative line graph sprouts like a pigtail out of the center of the first big bar.
 Figure 2: Typical Pareto chart |
I remember wondering, “Everyone calculates and draws control charts the same way; why is there such a vast difference in the way Pareto charts are drawn?”
Fast forward to present time. Recently, my friend Jack ReVelle sent me a copy of the training materials he used at Hughes Aircraft during the 1980s. On the cover is a Pareto chart in the same style as I was taught. This made me wonder: “Who invented the Pareto chart?” Since unanswered questions are often the predecessors to discovery, I decided to research the question of where the Pareto chart—as I knew it—originated. To do this, I needed to step back to the chart’s namesake, Vilfredo Pareto, himself.
Vilfredo Pareto was an Italian mathematician and economist. In the late 1800s, he created what we've come to know as the 80–20 rule, which observes that 20 percent of the population of Italy owned 80 percent of the wealth. Pareto’s work helped inspire Joseph M. Juran, who laid the groundwork for total quality management (TQM) and Six Sigma.
In Juran on Quality by Design (Free Press, 1992) pp. 68–71, Juran talks about the expansion of Pareto’s 80–20 rule to encompass fields other than economic distribution, describing it as “The Non-Pareto Principle: Mea Culpa.” Whereas Pareto’s application of the 80–20 rule only applied to wealth, Juran generalized it to many other phenomena, especially defects.
Like Pareto, Juran found that “quality defects are unequal in frequency.” In describing this, he often spoke of the “vital few and the trivial many,” and later, the “vital few and the useful many.” Juran named this pattern the “Pareto Principle” and later said that he could have called it the “Juran Principle.”
Because the terms “Pareto chart” and “Pareto Principle” seem to have gotten tangled up with Pareto analysis, many people jump to the conclusion that the creator of the Pareto chart was Juran as well.
However, only one Pareto chart with cumulative line graph appears in the Juran Handbook, Fourth Edition (McGraw-Hill, 1988) pg. 22.54. In addition, no listing for “Pareto chart” or “Pareto diagram” appears in the subject index. Instead, Juran used sorted bar/column charts (pp. 21.17, 22.21) to show the Pareto principle (figure 3). Because of this, it’s my contention that while Juran is unquestionably the origin of the phrase “Pareto Principle” as well as the first to apply it to other phenomena, he is not the creator of the Pareto chart as we know it today.
 Figure 3: Bar charts showing the Pareto principle |
A Pareto chart combines a bar graph with a cumulative line graph. Juran admits that the first edition of the Quality Control Handbook (McGraw-Hill, 1951) should have identified Max Otto Lorenz as the developer of the cumulative curve (Juran on Quality by Design, pp. 68–71). The exponential curve of wealth distribution was described by Lorenz in “Methods of Measuring the Concentration of Wealth,” American Statistical Association, vol. 9 (1904–19055), pp. 200–219. But Lorenz’s curve is neither a Pareto chart nor the cumulative line graph.
So if Juran created and expanded the Pareto Principle but didn’t create the Pareto chart, who did?
The concepts within Juran’s 1951 edition of Quality Control Handbook found an eager audience with Japanese manufacturers, and Juran was invited by the Japanese Union of Scientists and Engineers to visit Japan in 1952. Juran first visited Japan in 1954, meeting with manufacturing executives and lecturing at universities, and would return several more times during his lifetime.
Kaoru Ishikawa joined the Japanese Union of Scientists and Engineers in 1949 and worked to translate Juran’s works (as well as works by W. Edwards Deming) and integrate their ideas into Japan’s manufacturing sector.
In Ishikawa’s Guide to Quality Control (Asian Productivity Organization, 1986) pp. 42–49, we find clear examples of Pareto charts. (Note: Ishikawa called it a Pareto diagram.) Figure 4 shows an example from the book, figure 5.1 from p. 43, drawn with QI Macros.
 Figure 4: Ishikawa’s Pareto diagram |
The cumulative line begins at the x/y axis and exits the first bar at its top right corner, showing the cumulative effect of the bar. Additional points end at the edge of each subsequent bar. This enables users to “read” the percentages, even if they aren’t shown (figure 5).
 Figure 5: Additional points end at the edge of each subsequent bar, enabling users to “read” the percentages, even if they aren’t shown. |
Based on Ishikawa’s efforts during the early 1950s to translate Juran’s writings (which do not display a Pareto diagram) and the clarity of the presentation in his 1986 book, I believe that Ishikawa is the creator of not only the Ishikawa diagram (also known as a fishbone diagram), but also the sorted bar graph with a cumulative line chart that we now call the Pareto chart.
The 4-50 Rule
In my work during the last 25 years, I’ve noticed that Pareto’s Rule, or the Pareto Principle, is a power law. This means that the 80–20 distribution applies not only to the analysis as a whole, but also within the Pareto chart bars. As little as 4 percent of any process causes more than 50 percent of the waste, rework, and lost profit. I call this the 4–50 Rule (figure 6). The practical application of this rule tells us that only one process step, out of every 25, causes half of the waste, rework, and lost profit.
 Figure 6: As little as 4 percent of any process causes more than 50 percent of waste, rework, and lost profit. |
Summary
Though calling the “Pareto Principle” the “Juran Principle” might have been more accurate, I do think that the alliteration of “Pareto Principle” played some part in the popularity of the concept. I’m grateful to Juran for bringing it to our attention. Whether the Pareto chart was created by Ishikawa (as I suspect), Juran, or someone else, I’m grateful to whomever created it. This tool is essential to quality improvement in all industries.
If anyone has any additional insight into the origin of the Pareto chart, it would be an excellent addition to the history of quality.
Quality Digest does not charge readers for its content. We believe that industry news is important for you to do your job, and Quality Digest supports businesses of all types.
However, someone has to pay for this content. And that’s where advertising comes in. Most people consider ads a nuisance, but they do serve a useful function besides allowing media companies to stay afloat. They keep you aware of new products and services relevant to your industry. All ads in Quality Digest apply directly to products and services that most of our readers need. You won’t see automobile or health supplement ads.
Our PROMISE: Quality Digest only displays static ads that never overlay or cover up content. They never get in your way. They are there for you to read, or not.
So please consider turning off your ad blocker for our site.
Thanks,
Quality Digest
Comments
Joseph Juran and the Pareto Plot
Extremely interesting paper by Jay Arthur, but the third edition (1974) of Juran's Quality Control Handbook has on page 2-18 (Figure 2.5, 'Pareto analysis of weaving imperfects', a bar chart, cumulative line type Pareto chart that looks pretty much like the modern Pareto chart except that the line starts at the top of the first bar, and the y axis continues to 100%. This chart is clearly listed in the Subject Index on page 53, as 'Pareto diagram, 2-18, 2-19'. Interesting to see what is the seond (1962) edition as Leland Wilkinson, in Revising the Pareto Chart. The American Statistician, 60:4, 332-334, 2006, who pointed out that the chart did not appear in the first (1951) edition, but did in later editions. Wilkinson actually prefers the revised, cumulative barchart form, which Juran 1974 calls the 'Alternative Pareto diagram' on page 2-19, Figure 2-6.
So, the (Pareto) plot thickens?!
Who Invented the Pareto Chart
The kernel of this idea has a rule of thumb application in many places, but it is commonly misused, according to one of the JUSE Counselors who worked with us called Dr. Shiba. He stated, for example, it is a misuse to state a solution to a problem "fits the 80/20 rule" just because it fits 80% of the cases; it must also be that the solution requires only 20% of the resources that would be needed to solve all cases. Additionally, it is a misuse of the 80/20 rule to interpret a small number of categories or observations.
According to Dr Shiba, this is a special case of the wider phenomenon of Pareto distributions. If the Pareto index α, which is one of the parameters characterizing a Pareto distribution, is chosen as α = log45 ≈ 1.16, then one has 80% of effects coming from 20% of causes. Using the Shiba Method, It follows that one also has 80% of that top 80% of effects coming from 20% of that top 20% of causes, and so on.
The bottom line is this: 80 percent of 80% is 64%; 20% of 20% is 4%, so in the Shiba Method, this implies a "64/4" law; and similarly implies a "51.2/0.8" law. Similarly for the bottom 80% of causes and bottom 20% of effects, the bottom 80% of the bottom 80% only cause 20% of the remaining 20%. This is broadly in line with the world population/wealth table discussed elsewhere, where the bottom 60% of the people own 5.5% of the wealth, approximating to a 64/4 connection.
Shiba's Law: the 64/4 correlation, also implies a 32% 'fair' area between the 4% and 64%, where the lower 80% of the top 20% (16%) and upper 20% of the bottom 80% (also 16%) relates to the corresponding lower top and upper bottom of effects (32%). This is also broadly in line with the world population table, where the second 20% control 12% of the wealth, and the bottom of the top 20% (presumably) control 16% of the wealth.
Confusing but amusing to some.