The ninety-fifth article examined a chart whose numbers nobody could trace. This one examines a rule whose author traced it himself, published the correction, and watched it change nothing.
Key Takeaway
In a paper titled "The Non-Pareto Principle; Mea Culpa", Joseph Juran writes: "I was forced to confess that I had mistakenly applied the wrong name to the principle. This confession changed nothing." He adds that Pareto's "models were not intended to be applied to other fields."[1] Our own arithmetic: on an invented firm, dropping the bottom eighty percent of customers moves profit from zero to negative $60,000, because their contribution was covering overhead that stayed behind.
The Verdict, Stated First
Five claims, in descending order of confidence.
One. The man who named the principle published a paper saying he named it wrongly, which we obtained in full, and he states plainly that the universal was his own rather than Pareto's.
Two. He never states the ratio as eighty to twenty. His formulation is qualitative throughout, and the numbers were attached by others.
Three. Eighty-twenty is one parameter value of a distribution, not a law, and on our own arithmetic it corresponds to a shape parameter of about 1.161 with nothing to distinguish it.
Four. On our own arithmetic the rule has no stopping point, because the argument for the first cut applies with identical force to the survivors, indefinitely.
Five. And on our own arithmetic the rule is silent on the number the decision actually turns on, which is the fixed cost to serve a customer.
The fifth is the one a business owner should take away, ours, and it is the reason this article is worth eight thousand words rather than a correction.
Two of those five are historical and three are arithmetic, ours, which is a deliberate balance. The arithmetic would stand even if every historical claim here were overturned, and we have built the article so a reader can check that for themselves.
Our Grades For These Claims
Applying the scheme from the first article in this series. The sourcing here is the second strongest we have had, after the previous article's opposite.
Grade A-plus for Juran's own account, obtained in full as a paper hosted by the institute that carries his name.
Grade A for our own arithmetic, all of which is checkable and none of which depends on any historical claim.
Grade C for what Pareto himself published, which we did not obtain and which reaches us through secondary accounts that disagree about the year.
Grade D for the claim that Italian land was eighty percent owned by twenty percent of the population, which we have seen asserted repeatedly and have not verified anywhere near the source.
That last grade matters more than it looks, ours. The single most repeated fact about the Pareto principle is the one we could not check, and we decline to assert it.
Four sources here repeat it and none of them cites a page, ours. That is the pattern a claim shows when everybody is copying rather than reading, which is the ninety-fifth article's finding arriving in a much better-sourced article.
A Note On Method
Everything here is verified to August 2026.
We obtained Juran's 1974 paper in full[1], hosted as a PDF by the institute he founded, and every quotation attributed to him comes from it.
We did not obtain any work by Vilfredo Pareto, in any language or edition, and we say almost nothing about what he wrote beyond what Juran reports.
Secondary accounts of Pareto's observation come from an encyclopedia[2] and from trade and consultancy sources[3][4], all flagged, and they disagree with each other about the date.
All arithmetic is ours. The distribution mathematics is standard and the business figures are invented; every conclusion drawn from them is our own construction.
This article discusses a management heuristic and is not accounting, tax or business advice. A firm considering changes to its customer base should work from its own numbers with its own advisor.
The Claim As Used
What the rule does in practice, before we look at what it says.
A consultancy states it as: "80% of effects arise from 20% of the causes", glossed as twenty percent of your activities accounting for eighty percent of your results[4]. A consultancy, flagged.
Four observations, ours.
In a small business it arrives as a licence to cut. Twenty percent of clients, products or activities matter, so the remainder can go.
It is unusually portable, which is its appeal and its problem. The same sentence is applied to customers, products, defects, sales staff and hours in a week, with no argument that these behave alike.
And it arrives with two forms of authority stacked. A named economist and a specific pair of numbers, which together make it sound measured.
Both are worth examining, and the first one comes apart in a document written by the man who put it there.
One note on why the stacking matters, ours. A named economist supplies credibility and a pair of numbers supplies precision, and neither would carry the rule on its own.
Mea Culpa
The document, which we obtained in full and which is unusual in the same way the ninety-fourth article's was.
It is titled "The Non-Pareto Principle; Mea Culpa", dated 1974, and hosted by the institute the author founded[1].
He writes: "Years ago I gave the name 'Pareto' to this principle of the 'vital few and trivial many.' On subsequent challenges, I was forced to confess that I had mistakenly applied the wrong name to the principle. This confession changed nothing, the name 'Pareto principle' has continued in force, and seems destined to become a permanent label for the phenomenon."[1]
Four observations, ours.
The confession was not new even in 1974. He cites an earlier correction of his own from 1950[1], so by the time of this paper he had been correcting it for a quarter of a century.
The phrase "This confession changed nothing" is the sentence this whole series keeps arriving at, and here the author says it about himself.
His stated motive is specific: "to minimize this tendency to embroider the work of a distinguished Italian economist"[1], since he found contemporary authors attributing to Pareto "additional things which he did not do."
And he is unsparing about the mechanism of his own error, which we come to below. It was a caption.
Two things make this document unusual as evidence, ours. It is a statement against interest, since nothing in it helps the author, which is the same reason the ninety-fourth article trusted its central source.
And it is self-archived by the institution that carries his name, which means the organisation with the most commercial interest in the principle is hosting the correction to it.
What He Actually Claimed
The principle in its author's own words, which are more careful than the version in circulation.
He defines it as "a shorthand name for the phenomenon that in any population which contributes to a common effect, a relative few of the contributors account for the bulk of the effect."[1]
On origin he is direct: "Where then did the universal originate? To my knowledge, the first exposition was by myself." And: "Had I been structured along different lines, assuredly I would have called it the Juran principle. However, I was not structured that way. Yet I did need a shorthand designation, and I had no qualms about Pareto's name."[1]
Four observations, ours.
He describes the error's origin precisely. In the first edition of his handbook, a caption read that Pareto's principle of unequal distribution applied to both wealth and quality losses, and "the caption implies that he had generalized the principle of unequal distribution into a universal. This implication is erroneous."[1]
So the misattribution entered through a caption under a pair of curves, which is a smaller thing than most of the errors this series has examined and travelled further than any of them.
He also notes that the challenges came from named colleagues and that they "forced me to do what I should have done in the first place, to inform myself on just what it was that Pareto had done"[1].
And we would note the plain honesty of that sentence. He does not say the criticism was unfair or the distinction pedantic, which is how such challenges are usually received.
He also names the mechanism precisely enough to be checked, ours. A caption is a place where precision routinely lapses, because it is written last, read first, and reviewed by nobody.
He Never Used The Numbers
Our own observation, from reading the paper, and we think it is the most useful thing in this article.
Four observations.
The numbers eighty and twenty do not appear in Juran's statement of the principle. His formulation is "a relative few of the contributors account for the bulk of the effect"[1], which is qualitative.
"A relative few" and "the bulk" are the operative terms, and neither is a quantity. The claim is about shape, not proportion.
Stated that way, the principle is far more defensible and far less useful. That contributions are unequally distributed is nearly always true; that the split is precisely eighty to twenty is a specific empirical claim about a specific population.
And that is the substitution the business version makes. A qualitative observation about inequality acquired a pair of numbers, and the numbers are what licence a decision.
Two consequences follow, ours. The principle as its author stated it cannot be wrong in any interesting way, because "a relative few" and "the bulk" accommodate almost any degree of concentration.
And the version that can be wrong is the version nobody attributes to anybody, since the numbers have no author at all.
Not Intended For Other Fields
The sentence that settles the scope question.
Juran writes that examining Pareto's work "made clear to me what I had seen only dimly, that Pareto's work had been in the economic sphere and that his models were not intended to be applied to other fields."[1]
Four observations, ours.
This is a scope condition stated by the person who violated it, which makes it unusually authoritative.
The ninety-third article found the identical structure. A model fitted in one domain was extrapolated to another, and the extrapolation is where the trouble entered.
Juran's defence of the generalisation is worth stating fairly, because he does not abandon it. He maintains that the universal exists and that he was the first to write it down, and he says so in the same paper.
So the correct summary is narrower than "the Pareto principle is false." It is that the principle is Juran's, is qualitative, and borrowed a name and an air of mathematical authority from work that did not license the borrowing.
The Curves Were Not His Either
A second misattribution the same paper discloses.
Juran writes: "To make matters worse, the cumulative curves used in Quality Control Handbook, First Edition, should have been properly identified with Lorenz."[1]
He cites Lorenz, M.O., Methods of Measuring the Concentration of Wealth, American Statistical Association Publication, Vol. 9 (1904–1905), pp. 200–219[1].
Four observations, ours.
So the diagram and the name came from two different people, neither of whom was the source of the universal being illustrated.
The Lorenz curve is a genuine and still-used tool for measuring concentration, and it measures rather than asserts, which is the distinction the ninety-second article drew about diagrams.
We did not obtain the Lorenz paper and say nothing about its contents.
And we note the shape of the whole episode. A real measurement tool, a real economist's name, and a real observation about defects were combined into something none of the three supported, by a careful person who then spent decades saying so.
From Trivial To Useful
A revision that anticipates this article's arithmetic by fifty years.
The summary attached to the paper records that Juran coined the phrase "vital few and trivial many," later changed to the "useful many."[1]
An encyclopedia gives the reason: he preferred the later phrasing to dissuade an interpretation of the principle as the contribution of the eighty percent being without value[2]. An encyclopedia, flagged.
Four observations, ours.
The change is small and the intent is unmistakable. He saw people reading "trivial many" as "these do not matter" and changed the word.
It did not work, which is why the phrase in circulation is still the original one, and why the business application is still a licence to cut.
And the arithmetic below is, in effect, a quantitative demonstration of the point he made in a word choice. The many are not trivial, and we can show by how much.
We would note that this is the third consecutive article in which the originator's own qualification failed to travel. A caption travelled and two corrections did not, over fifty years, by the same author.
Eighty Is A Parameter
Our own arithmetic, and it reframes what kind of thing the rule is.
For a Pareto distribution with shape parameter a, the top p fraction of a population holds p raised to the power (a−1)/a of the total. That relationship is standard mathematics and not our invention.
Setting the top twenty percent to hold eighty percent and solving gives an exponent of about 0.1386, which corresponds to a shape parameter of about 1.161.
Four observations.
Eighty-twenty is therefore not a law. It is one value of one parameter, and nothing in the mathematics distinguishes 1.161 from any neighbouring value.
The rule is stated as though it were a general property of populations. It is a property of populations with one particular degree of concentration, and other degrees are equally possible.
Which means the rule contains a hidden empirical claim. Every use of it asserts that your business has this specific shape parameter, and nobody who uses it has measured that.
And the claim is checkable in an afternoon, which we come to below.
One further implication worth stating, ours. Because the parameter is estimable, a firm can find out where it actually sits, which converts the rule from an assertion into a measurement and is the whole remedy.
What Nearby Values Give
Our own arithmetic, showing how sensitive the famous split is.
At a shape parameter of 1.05, the top twenty percent holds 92.6 percent. At 1.10, 86.4 percent. At 1.161, 80.0 percent. At 1.25, 72.5 percent. At 1.50, 58.5 percent. At 2.00, 44.7 percent. At 3.00, 34.2 percent.
Four observations.
A shift in the parameter from 1.161 to 1.5, which is not a large change, moves the top twenty percent's share from eighty to fifty-nine.
At a parameter of 2, the top fifth holds under half. The population is still unequal and the rule as stated is simply wrong about it, which is the situation many real businesses are in.
And in the other direction the concentration gets extreme fast. At 1.05 the top fifth holds ninety-three percent, which would make the case for cutting far stronger than the rule claims.
So the rule is not conservative or cautious. It is a single point on a continuum, quoted as though it were the continuum, and a firm on either side of it is being told the wrong thing.
And the direction of the error matters commercially, ours. A firm less concentrated than eighty-twenty will overestimate how much it can afford to lose, which is the more dangerous of the two mistakes.
The Sixty-Four Percent Claim, Checked
A claim we have seen made often, and we checked whether it follows. Our own arithmetic.
People commonly say that the rule applies to itself: the top four percent holds sixty-four percent, being eighty percent of eighty percent.
Working the distribution formula through: the top 20 percent holds 80.0 percent. The top 4 percent holds 64.0 percent. The top 0.8 percent holds 51.2 percent. The top 0.16 percent holds 41.0 percent.
Four observations.
The claim is exactly correct, and the values are precisely 0.8 raised to successive powers. We expected to find an error here and did not.
That said, it holds only because the distribution is scale-free by assumption. Self-similarity is a property we put into the model, so recovering it is not a discovery.
Real customer bases are bounded, in a way the mathematics is not. A firm with forty clients cannot have a meaningful top 0.16 percent, and the recursion runs out of population long before it runs out of arithmetic.
And that boundedness is the first hint of the problem in the next section, which is what the recursion licenses.
Two practical notes on the self-similarity, ours. It is why the rule feels so robust, since applying it at any scale returns something that looks like a confirmation.
And that is a warning rather than a comfort. A model that confirms itself at every scale is not being tested at any of them.
The Rule Has No Stopping Point
Our own arithmetic, and we think it is the decisive structural objection.
Suppose a firm accepts the rule and cuts the bottom eighty percent of customers. It keeps twenty percent of customers and eighty percent of revenue.
If the distribution is scale-free, the survivors are themselves distributed eighty-twenty, and the identical argument applies again with identical force.
Running the cut repeatedly: after one round, 20 percent of customers and 80 percent of revenue. After two, 4 percent and 64. After three, 0.8 percent and 51. After four, 0.16 percent and 41. After six, 0.0064 percent of customers and 26 percent of revenue.
Four observations.
Nothing in the rule says when to stop. The reason for the first cut is the reason for the sixth, unchanged, which means the rule is a gradient rather than a decision.
A defender would say that obviously you stop after one round, and that is exactly the point. The stopping decision comes from somewhere else, and wherever it comes from is doing the actual work.
The recursion also exposes what the first cut costs. Each round discards a fifth of remaining revenue, which is a large sacrifice to be justified by a rule that cannot say why one round is right.
And the ninety-second article's finding applies here too. A framework that gives an ordering does not thereby give a threshold, and the threshold is the entire commercial question.
One honest concession to a defender, ours. Most heuristics have this property, and it does not make them useless; it makes them prompts rather than rules.
The trouble is that this one is used as a rule, and its name and its numbers are why.
The Decision The Rule Cannot Make
Our own arithmetic on an invented firm. Every figure below is ours.
One hundred customers, one million dollars of revenue, split exactly eighty-twenty. The top twenty bill $40,000 each; the bottom eighty bill $2,500 each. Variable cost runs at 70 percent of revenue for both groups.
The only quantity we vary is the fixed cost to serve one customer, which the rule does not mention.
At $200 per customer: the bottom eighty produce $44,000 of profit. At $500: $20,000. At $1,000: negative $20,000. At $1,500: negative $60,000.
Four observations.
The decision reverses inside that range, and the eighty-twenty split is identical in every row.
So two firms with exactly the same revenue concentration should do opposite things, and the rule cannot tell them apart because it does not contain the quantity that separates them.
The figures are invented and the structure is not. Fixed cost to serve is a real quantity every business has, comprising onboarding, account management, billing, support and the share of any per-customer system licence.
And it is measurable. A firm can estimate it in an afternoon and the rule cannot supply it, which makes the rule strictly less useful than the work it displaces.
Two components most firms forget, ours. The owner's own time spent on a small account is a real cost, and it is usually excluded because nobody invoices for it.
And per-seat or per-client software licences scale with customer count, so a firm on modern subscription tooling has a higher fixed cost to serve than it did a decade ago.
Where It Breaks Even
The exact tipping point, which is cleaner than we expected. Our own arithmetic.
A small customer billing $2,500 at a seventy percent variable cost contributes $750 before fixed costs.
So the break-even fixed cost to serve is exactly $750 per customer per year. Below it the bottom eighty percent make money; above it they lose money.
Four observations.
That figure is not a coincidence and not our discovery. It is simply the contribution margin, and the whole decision is whether fixed cost to serve exceeds it.
Which is the ordinary, unglamorous answer that any management accountant would give, and it makes no reference to eighty or twenty at all.
The rule's contribution to this decision is precisely nothing, and its cost is that it supplies an answer with confidence before the question has been asked.
And the correct procedure is available to any firm. Compute contribution per customer, estimate fixed cost to serve, and compare the two, which is a morning's work and gives an answer about your business rather than about a distribution.
We would add one refinement, ours. Do it per customer rather than per group, because the bottom eighty percent is not homogeneous and will contain both the worst accounts and some perfectly good small ones.
The Overhead Does Not Leave
The trap, and it is the most expensive item in this article. Our own arithmetic on an invented firm.
Take the same hundred customers and one million of revenue. Variable costs of $700,000, and overhead of $300,000 that does not fall when customer count falls: premises, systems, insurance, the owner's time.
The firm is at break-even, with profit of zero.
It cuts the bottom eighty customers. Revenue falls to $800,000 and variable costs fall to $560,000. Overhead stays at $300,000.
Profit is now negative $60,000.
Four observations.
The eighty customers were contributing $60,000 toward overhead, which is exactly the amount by which profit fell.
They will frequently have looked unprofitable in a management report, because any allocation of overhead across a hundred customers assigns them a share they cannot cover, and allocated overhead is not avoidable cost.
This is the oldest trap in cost accounting and it is still routine. The relevant test is contribution, not allocated profit, and the two give opposite answers precisely for small customers.
And the rule makes it worse rather than causing it. It supplies a respectable-sounding reason to act on a number that was already misleading, which is how a heuristic does its damage.
Two further costs the arithmetic does not capture, ours. Small customers grow, and a firm that cuts its smallest accounts has removed its own pipeline.
And capacity does not shrink on the day the customers leave. Staff hired to serve a hundred accounts are still employed the following month, so the saving arrives later than the revenue loss, if at all.
Concentration Is A Risk, Not A Result
An inversion worth making explicit, because the rule presents a warning as a discovery. Ours.
Four observations.
A firm where the top fifth of customers really does supply eighty percent of revenue has a small number of relationships it cannot afford to lose.
Any lender or acquirer reads that number the opposite way to the rule. High customer concentration is a standard adverse finding in diligence, and it reduces what a business is worth.
The arithmetic is unforgiving. Losing one customer out of twenty who each supply four percent of revenue costs four percent, and there is no version of that which a firm absorbs comfortably.
So the same fact supports two opposite readings. The rule reads concentration as showing where the value is; a buyer reads it as showing where the fragility is, and the second reading is the one that gets priced.
The Customers You Have Not Met Yet
The dynamic problem, which no snapshot can see. Ours.
Four observations.
Every large customer was once a small one, in most businesses, and a ranking taken today cannot distinguish a small account that will stay small from one that will not.
The rule operates on a single period's revenue, which is the only thing on the report, and treats position in that ranking as a property of the customer rather than of the moment.
A firm can test this cheaply on its own history. Take the ranking from three years ago and see where today's largest accounts sat, which takes an hour and answers the question directly.
And if a meaningful share of today's top customers were in the bottom eighty percent three years ago, the rule would have instructed you to fire them, which is the sharpest available test of whether to follow it.
We would note that this test can come out either way, ours, and that is what makes it worth running. A firm whose large accounts arrived large has a genuinely different situation from one that grew them, and only its own records can say which it is.
What Actually Survives
Our reading, stated directly.
Five statements.
The universal is Juran's, not Pareto's, on Juran's own account, and he says the first exposition was his.
His formulation contains no numbers, describing a relative few contributors accounting for the bulk of an effect.
Pareto's models were not intended for other fields, in Juran's words after he went and checked.
On our own arithmetic, eighty-twenty is one parameter value among many, and neighbouring values give substantially different splits.
And on our own arithmetic, the rule cannot make the decision it is used for, because that decision turns on fixed cost to serve and on avoidable versus allocated overhead, neither of which the rule contains.
Those five are what we would defend, ours, and the first three rest on a document we obtained in full from the man who named the thing, which is the second time in three articles this series has been in that position.
The Rule Is Not Useless
Because this could read as a dismissal, and that would be wrong. Ours.
Four observations.
Contributions really are unequally distributed, in most populations most of the time, and that is worth knowing and is frequently forgotten.
Juran's original application was to defects, where the logic is much stronger. Fixing the most common defect first is nearly always right, because defects have no revenue to lose.
That asymmetry is the key, ours. The rule works well where the tail is pure cost and badly where the tail is revenue, and the business version applies it mostly to the second case.
And as a prompt it is genuinely useful. Asking which of your customers or products account for most of your revenue is a good question, and the rule's failure is in answering it before you look.
Which is the fairest summary we can give, ours. It is a good question wearing the costume of an answer, and stripping the costume off leaves something worth keeping.
The Question It Replaces
What gets skipped, ours, and it is the general lesson.
Four observations.
A heuristic's cost is the enquiry it prevents. The rule supplies a distribution, a conclusion and an action, so nobody computes the distribution.
Here the enquiry is cheap. Sorting your customers by revenue and computing the top twenty percent's share takes minutes, and gives you your number rather than somebody's.
The ninety-third article made the same point about a different figure. A benchmark should be a starting point and is routinely used as a threshold, and the difference is whether anybody checked.
And the check has a second benefit. Knowing your actual concentration tells you about your risk, since a firm where the top twenty percent really do supply eighty percent has a dependency problem the rule presents as a virtue.
Measure Your Own Concentration
The practical procedure. Ours, and not accounting or business advice.
Four steps.
Sort customers by revenue and compute the cumulative share. Read off what the top twenty percent actually holds, which may be forty percent or ninety.
Compute contribution, not allocated profit, for each group: revenue minus the costs that would actually disappear if the customer did.
Estimate fixed cost to serve one customer, which is the number the whole decision turns on and which nobody has on a report.
And separate avoidable overhead from allocated overhead before concluding anything, because the second does not leave when the customer does.
One test for that separation, ours, and it is simple. Ask what cheque stops being written if the customer leaves tomorrow. If no cheque stops, the cost was allocated rather than caused.
The Other Uses
Where the rule is applied besides customers, and how the analysis changes. Ours.
Four observations.
Applied to defects or complaints, it is at its strongest, because the tail is cost and eliminating it has no revenue consequence.
Applied to products, it is weaker than it looks, because slow-moving lines frequently support the fast ones by making a firm worth visiting, and that interaction is invisible to a ranking.
Applied to time management, it is nearly meaningless as stated. There is no natural unit of activity to rank, so the twenty percent can be defined after the fact to contain whatever worked.
And applied to staff, it is the most dangerous of the four, since the ranking is noisier than anybody admits, which the eighty-eighth and eighty-fourth articles both examined.
One property distinguishes the strong case from the weak ones, ours, and it is worth naming. Defects do not respond to being ranked. Customers, products and staff do, and a ranking that changes the thing it measures is a different instrument entirely.
Bibliographic Note
The series keeps a count, and this article produced three, one of them a four-way disagreement.
Pareto's observation is dated 1895 by one consultancy[4], 1896 by one commentary[3], and 1906 by an encyclopedia and a trade publication[2][5]. A physics paper's reference list gives the book as published in 1897[6].
Juran's paper is dated 1974 in the institute's own header and filename[1], and is commonly cited elsewhere as a 1975 journal article.
And an encyclopedia dates Juran's development of the concept to 1941[2], where his own narrative places the encounter with Pareto's work in the late 1930s and the recognition of the universal in the late 1940s[1].
Three observations, ours.
Four different years for one observation is the widest spread we have recorded, and it is fitting that the article about a misattributed principle should produce it.
The 1941 discrepancy is the one that matters, since a secondary source is contradicting the man's own account of his own career, which is a reminder about where to look.
We hold that lightly in one respect, ours. A recollection written decades later can be wrong about its own dates, so the discrepancy establishes a conflict rather than an error.
That brings the running count of bibliographic variants across this series to forty-seven.
Three in one article is our highest single-article total, ours, and every one of them concerns a date rather than a page number, which is a category we had not seen cluster before.
What To Do
Stop citing Pareto. The universal is Juran's on his own account, he said so repeatedly from 1950 onward, and Pareto's models were not intended for other fields.
Drop the numbers. Juran's formulation is that a relative few contributors account for the bulk of an effect, and it contains no eighty and no twenty.
Compute your own concentration. Sorting customers by revenue takes minutes and tells you whether your top fifth holds forty percent or ninety.
Find your fixed cost to serve. On our own arithmetic that number, not the concentration, decides whether small customers are worth keeping.
Use contribution, not allocated profit. Allocated overhead makes small customers look unprofitable when they are covering costs that will not leave with them.
Notice that the rule has no stopping point. The argument for cutting the bottom eighty percent applies unchanged to the survivors, six times over.
Apply it to defects rather than to revenue. The logic is strongest where the tail is pure cost, which is what Juran was working on when he found it.
And treat high concentration as a risk rather than a finding. A firm whose top fifth really does supply eighty percent of revenue has a dependency, not a strategy.
The Limits Of This Analysis
Several caveats matter. This article discusses a management heuristic and is not accounting, tax or business advice; a firm considering changes to its customer base should work from its own figures with its own advisor. Everything is verified to August 2026. We did not obtain any work by Vilfredo Pareto, in any language or edition, so everything here about what he wrote comes either from Juran's account of examining it or from secondary sources that disagree with each other about the year by as much as eleven years. We did not verify the claim that eighty percent of Italian land was owned by twenty percent of the population, which is the single most repeated fact about this principle, and we decline to assert it. We did not obtain the Lorenz paper Juran cites, nor his 1950 correction, nor either edition of his handbook, so his account of his own error is uncorroborated by the documents it describes. Juran's paper is self-archival, hosted by the institute he founded, and is a first-person recollection written decades after the events; we treat it as credible principally because it is a statement against interest, and a reader who weighs it differently would reach different conclusions. All arithmetic is ours. The distribution mathematics is standard, but every business figure is invented: the hundred customers, the million in revenue, the seventy percent variable cost, the fixed costs from $200 to $1,500, and the $300,000 of overhead. A firm with different figures gets different answers, and the break-even of $750 is simply the contribution margin of our invented small customer. Our recursion argument assumes the distribution remains scale-free after each cut, which real bounded customer bases do not, so the six-round figure is a demonstration of the rule's logic rather than a prediction. And our claim that the rule works better for defects than for revenue is our own reasoning, not a finding of any source.
Frequently Asked Questions
Did Pareto come up with the 80/20 rule?
Where do the numbers eighty and twenty come from?
Is the claim that the top four percent holds sixty-four percent correct?
Should I fire my smallest customers?
What is the overhead trap?
Why does the rule have no stopping point?
Is the principle ever useful?
References
- Juran, J. M. (1974), The Non-Pareto Principle; Mea Culpa, from the archives of the Juran Institute, hosted as a PDF on the institute's own site. Obtained in full. States that the Pareto principle is a shorthand name for the phenomenon that in any population which contributes to a common effect, a relative few of the contributors account for the bulk of the effect; that years ago the author gave the name Pareto to this principle of the vital few and trivial many; that on subsequent challenges he was forced to confess that he had mistakenly applied the wrong name to the principle, citing his own earlier correction in Industrial Quality Control, October 1950; that this confession changed nothing and the name has continued in force; that contemporary authors have fabricated embellishments and attributed to Pareto additional things which he did not do; that his motive in offering the paper is in part to minimize this tendency to embroider the work of a distinguished Italian economist; that he first observed the unequal frequency of quality defects as a young engineer in the mid-1920s; that he encountered Pareto's work through a General Motors executive salary study; that by the late 1940s he had recognised the principle as a true universal; that in preparing the first edition of the Quality Control Handbook he wrote a caption reading that Pareto's principle of unequal distribution applied to distribution of wealth and to distribution of quality losses; that the caption implies Pareto had generalised the principle into a universal and that this implication is erroneous; that the Pareto principle as a universal was not original with Pareto; that to his knowledge the first exposition was by himself; that had he been structured along different lines he would have called it the Juran principle, but that he needed a shorthand designation and had no qualms about Pareto's name; that challenges from named colleagues forced him to inform himself on what Pareto had actually done; that this examination made clear that Pareto's work had been in the economic sphere and that his models were not intended to be applied to other fields; and that the cumulative curves used in the handbook should have been properly identified with Lorenz, citing Lorenz, M.O., Methods of Measuring the Concentration of Wealth, American Statistical Association Publication, Vol. 9 (1904–1905), pp. 200–219. An attached summary records that Juran coined the phrase vital few and trivial many, later changed to the useful many. Note: a self-archived first-person recollection hosted by the institute the author founded, NOT peer reviewed, and written decades after the events it describes. Our source for every statement attributed to Juran. We treat it as credible principally because it is a statement against interest. juran.com
- Encyclopedia entry on the Pareto principle, also known as the 80:20 rule, the law of the vital few and the principle of factor sparsity, stating that for many outcomes roughly 80 percent of consequences come from 20 percent of causes; that in 1941 Juran developed the concept in the context of quality control after reading the works of Vilfredo Pareto, who wrote in 1906 about the 80:20 connection while teaching at the University of Lausanne; that Pareto noted approximately 80 percent of Italy's land was owned by 20 percent of the population; and that later in his career Juran preferred to describe this as the vital few and the useful many, to dissuade from an interpretation of the principle as the contribution of the 80 percent being without value. Note: an encyclopedia, NOT an academic source, flagged at every use. Our source for Juran's reason for revising the phrase. Its date of 1941 conflicts with Juran's own narrative, and we record that as a bibliographic variant. en.wikipedia.org
- Commentary article on the Pareto principle in an organisational design publication, stating that in 1896 Pareto published a study in his book Cours d'economie politique observing that approximately 80 percent of Italy's land was owned by 20 percent of the population, that he noted similar patterns in other countries and extended the observation to income distribution, and that the principle as known today was popularised much later by Juran, who coined the terms vital few and trivial many in the 1940s while applying the idea to quality control. Note: an independent commentary site, NOT an academic source, flagged. Recorded principally for its date of 1896, which differs from three other sources used here. sergiocaredda.eu
- Consultancy guide to the Pareto principle, published by the institute carrying Juran's name, stating that the principle illustrates that 80 percent of effects arise from 20 percent of the causes, glossed as 20 percent of activities accounting for 80 percent of results; that it takes its name from Pareto, who observed that a relative few people held the majority of the wealth back in 1895; that Pareto developed logarithmic mathematical models and Lorenz developed graphs to illustrate them; and that Juran was the first to point out that what Pareto and others had observed was a universal principle. Also advises that where a Pareto diagram does not produce a clear picture of the vital few because categories are nearly equal, a team should try another way of classifying the problem and will almost certainly find a classification producing a vital few. Note: a consultancy's marketing and training material, NOT an academic source, flagged. Our source for the rule as stated in current commercial use, and for the date 1895. juran.com
- Trade publication article on the principle, stating that in 1906 Pareto created a mathematical formula to describe the unequal distribution of wealth in his country, observing that twenty percent of the people owned eighty percent of the wealth; that in the late 1940s Juran inaccurately attributed the 80/20 rule to Pareto, calling it Pareto's principle; that Juran recognised a universal principle he called the vital few and trivial many; and that in an early work a lack of precision on Juran's part made it appear that he was applying Pareto's observations about economics to a broader body of work. Note: a trade publication, NOT an academic source, flagged. Recorded for its independent confirmation of the misattribution and for its date of 1906. labmanager.com
- Physics preprint on Pareto's law for individual income and the debt of bankrupt companies, whose reference list gives the primary work as V. Pareto, Le Cours d'Economie Politique, Macmillan, London, 1897. Note: cited here ONLY for its bibliographic record of Pareto's book, which gives a fourth publication year and a different title form from the other sources used. We did not read the paper's own findings and make no use of them. arxiv.org
This article discusses a management heuristic and is not accounting, tax or business advice. No work by Vilfredo Pareto was obtained in any edition, and the frequently repeated claim about Italian land ownership was not verified and is not asserted here. Juran's paper is a self-archived first-person recollection written decades after the events. All arithmetic is the authors' own; the distribution mathematics is standard and every business figure is invented.