The Pareto principle, better known as the 80/20 rule, says that in any group contributing to a shared result, a small share of the contributors account for most of the result. Measured examples are everywhere: in 2019 the top 5% of people by healthcare spending accounted for 48.8% of all US healthcare expenditure, 1% of the Swedish population accounted for 63% of violent crime convictions, and across 238 consumer brands the top 20% of buyers produced 73% of dollar sales. The split is almost never exactly 80/20. What repeats is the shape, not the ratio.
“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 is the definition from the man who named the Pareto principle, and it is looser than the version most people carry around. Juran’s wording contains no 80, no 20, and no promise that the two numbers add up to anything in particular.
What the Pareto principle actually says
The Pareto principle claims that contributions to an outcome are unevenly distributed, and usually far more unevenly than people expect. Roughly 20% of the causes produce roughly 80% of the effect, the remainder produce the rest, and the exact figures move from case to case. Three things follow from that, and each one gets lost in the usual retelling.
- The two numbers are not a sum. One number is a share of causes, the other is a share of effects. They are measured on different quantities, so a 1 and 63 split, or a 20 and 73 split, is just as much a Pareto distribution as 20 and 80. A split that happens to add to 100 is a coincidence of the most famous example.
- The Pareto principle is a finding, not a mechanism. The Pareto principle describes what a measured distribution looks like. The Pareto principle does not explain why the distribution is lopsided, and it does not predict that a new situation will be lopsided in the same proportion.
- The Pareto principle only means something once you measure. Any claim of the form “80% of our results come from 20% of our work” is a testable statement about a specific dataset. Without the data behind it, the claim is a hunch wearing a percentage sign.
Where the 80/20 rule came from, and why the name is wrong
The 80/20 rule is named after Vilfredo Pareto, an Italian economist who studied the distribution of income and published a logarithmic law describing it in Cours d’économie politique (1896 to 1897). Pareto found that the number of people earning above a given income fell away in a regular curve, and that the same curve fitted country after country. What Pareto did not do is claim the pattern held outside economics, or state any 80/20 split.
The generalisation came from Joseph M. Juran, the quality engineer, who needed a short name for a pattern he had been seeing since the 1920s in defect counts, absenteeism and accident causes. Juran put it in the first edition of the Quality Control Handbook in 1951 and attached Pareto’s name to it. He regretted that later, and said so in print.
“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.”
In the same paper Juran states plainly that “the Pareto principle as a universal was not original with Pareto”, credits the cumulative curve to Max Lorenz rather than to Pareto, and says the reason he did not call it the Juran principle is simply that he was not built that way. He also revised his own phrase: the trivial many became the useful many, because calling four fifths of your work trivial turns out to be a licence to neglect it.
One footnote worth having, because it appears in textbooks and even in peer-reviewed papers: the story that Pareto noticed 20% of the pea pods in his garden held 80% of the peas has no source in Pareto’s own writing. Treat the pea pods as folklore. The income curve is the real history.
Five Pareto principle examples with the numbers checked
Below are five cases where the concentration has actually been measured and published, rather than asserted. Each one is unpacked in its own section underneath the table, and every figure links to the primary source rather than to a secondary write-up.
| Where | The measured split | Source (year) |
|---|---|---|
| US healthcare spending | Top 5% of people accounted for 48.8% of all spending. Reading the published curve at the 20% mark gives about 81%. | AHRQ Medical Expenditure Panel Survey, 2019 data |
| Software crashes | For Microsoft Word, the first quarter of all error reports came from 0.057% of the distinct crash signatures. | Glerum et al., SOSP 2009 |
| Brand revenue | Top 20% of buyers produced 73% of a brand’s dollar sales, averaged over 238 brands. | Kim, Singh and Winer, Marketing Letters 2017 |
| Violent crime convictions | 1% of the population accounted for 63% of all violent crime convictions over 32 years. | Falk et al., Soc Psychiatry Psychiatr Epidemiol 2014 |
| Household income | A regular logarithmic curve across several countries. No fixed ratio, which is the point. | Pareto, Cours d’économie politique, 1896 to 1897 |
Example 1: a fifth of people account for four fifths of healthcare spending
The cleanest published Pareto principle example comes from the Agency for Healthcare Research and Quality, which measures the concentration of US healthcare expenditure every year. In 2019, the top 1% of people ranked by spending accounted for 20.7% of all healthcare expenditure, the top 5% for 48.8%, and the bottom half of the population for 3.0%.
The published concentration curve gives the 20% figure directly: the bottom 80% of the population accounted for 18.6% of expenditure, so the top 20% accounted for the other 81.4%. That is a genuine, government-measured 80/20, on a dataset of more than $2 trillion in spending, and it is the closest thing to the textbook ratio in this list.
Two details make the healthcare case useful rather than merely neat. The first is that the concentration is driven by something concrete: 36.7% of the top group’s spending went on inpatient stays, against 0.2% for the bottom half. The second is that about 14% of the US population had no personal healthcare expenditure at all that year. A long flat tail of near-zero contributors is what produces a steep Pareto curve, in healthcare and everywhere else.
Example 2: software crashes are far steeper than 80/20
Microsoft’s crash reporting system produced the most extreme published Pareto distribution in this list, and it is the example that best kills the idea that the ratio is fixed. In Debugging in the (Very) Large, the Windows Error Reporting team reported how error reports from the Office 2010 internal technical preview distributed across distinct crash signatures, which they call buckets.
| Program | Share of buckets producing the first quarter of reports | Share of buckets producing the last quarter |
|---|---|---|
| Word | 0.057% | 91.7% |
| Outlook | 0.058% | 93.1% |
| PowerPoint | 0.106% | 91.4% |
| Excel | 0.227% | 88.4% |
Read the Word row again. A quarter of everything users hit came from less than a sixteenth of one percent of the distinct bugs, and the last quarter of reports was spread across 91.7% of them. Calling that 80/20 would badly understate it. The practical consequence is the one the paper draws: with three weeks of reports from 9,000 internal users, the team had already fixed bugs responsible for over 22% of all error reports.
The software case also shows the Pareto principle moving over time. The authors describe teams climbing down their own error curve: as the most frequent crashes get fixed, the distribution flattens, and the next round of fixes buys less. A steep Pareto curve is a temporary opportunity, not a permanent property of the system.
Example 3: brand revenue is 73/20, not 80/20
The claim that 80% of revenue comes from 20% of customers is the most repeated business version of the Pareto principle, and it has been measured properly. Kim, Singh and Winer used six years of A.C. Nielsen panel data covering roughly 100,000 US households, and computed, for each of 238 brands across 22 product categories, the share of dollar sales coming from the top 20% of buyers.
The average came out at 0.73, published in Marketing Letters in 2017, with a standard deviation of 0.07. Measured by volume rather than dollars it was 0.72. Category averages ranged from about 0.64 for detergents up to 0.89 for cigarettes, and the hedonic categories clustered at the top: cigarettes, light beer and soft drinks all concentrated harder than paper towels.
The finding that matters most for anyone running a business is not the headline number but the direction of the variation. Brands with a higher market share had a lowerPareto ratio. Coca-Cola sat at 0.79 and Pepsi at 0.80, while the smaller 7 Up and A&W sat at 0.71 and 0.70. Growth in this category comes from new buyers rather than from existing loyal ones, so the more successful a brand becomes, the less true the 80/20 story gets. A rule that weakens exactly as you succeed is a poor thing to build a strategy on without checking it first.
Example 4: 1% of a population, 63% of the convictions
The starkest published concentration in this list comes from criminology. Falk and colleagues followed the entire Swedish population born between 1958 and 1980 and linked every violent crime conviction recorded from 1973 to 2004. The result is in the paper’s title: 1% of the population accounted for 63% of all violent crime convictions (Social Psychiatry and Psychiatric Epidemiology, 2014).
A 1 and 63 split is the clearest possible demonstration that the two numbers in a Pareto distribution are independent. Nobody looking at this dataset would say the Pareto principle fails because 1 and 63 do not make 100. The concentration is the finding, and the concentration here is extreme.
The criminology case carries a warning as well as an illustration, and the warning transfers directly to business use. A steep concentration tells you where the volume sits. A steep concentration does not, by itself, tell you what to do about it, and it does not license conclusions about the people in the tail. Both cautions apply just as much when the tail is your own customers or your own colleagues.
Example 5: income, the case Pareto actually studied
The original Pareto principle example is income distribution, and it is worth returning to because of what it does not contain. In Cours d’économie politique Pareto fitted a logarithmic curve to income data from several countries and found the same shape in each: the number of people earning more than a given amount fell away in a regular, predictable way, with a small group at the top holding a large share of the total.
What Pareto produced was a curve with a parameter, not a slogan with two numbers. Different countries and different periods gave different steepness. The 80/20 phrasing was a later convenience, attached by Juran and then hardened into a rule by everyone who repeated it. The single most useful thing to take from Pareto’s own work is that the steepness is a variable you measure, not a constant you assume.
How to find the 80/20 in your own work
Applying the Pareto principle to your own work means running the measurement, not estimating the answer. The version people usually do, sitting down to decide which 20% of their tasks matter, produces a list of the work they already enjoyed. A real 80/20 analysis takes about thirty minutes and runs in four steps.
- Pick one outcome you can count. Revenue, signups, support tickets resolved, bugs reported, hours lost to rework. One number, over one fixed period that has already finished. An outcome nobody counted cannot be analysed retrospectively, so start with something your tools already record.
- List the contributors and their contribution. Customers and what each spent. Task types and how many hours each consumed. Error messages and how often each fired. Keep the list flat and specific: categories that are too broad will hide the concentration you are looking for.
- Sort by contribution, largest first, and cumulate. Add a running total and a running percentage of the whole. The cumulative column is where the Pareto principle shows up or fails to, and either answer is worth having.
- Read off where the curve bends. Find the point where the cumulative percentage crosses about three quarters. Everything above that line is your vital few. Do not force the cut to land at 20%: in the Microsoft data the useful cut was a fraction of a percent, and in the detergent category it was closer to a third.
The output is one short list and one decision. The decision is what to do with the extra capacity, which is usually to give the vital few more attention rather than to abandon the tail, since most of that tail is work that still has to happen. For the week-level version of this, our post on productivity tips for work covers what to do with the time you free up, and deep work strategies covers how to protect it.
How to build a Pareto chart
A Pareto chart is the tool Juran actually gave people, and it is the honest test of whether the Pareto principle holds in your data. A Pareto chart is a bar chart of causes sorted from most frequent to least, with a cumulative percentage line drawn over the top. Where the line flattens is where the vital few end.
The chart needs five columns and nothing else. Copy this into a spreadsheet, fill in the first two columns, and let the last three calculate:
Cause Count % of total Cumulative Cumulative % ----------------------- ------- ----------- ----------- ------------ <cause with most count> <count> <count/sum> <count> <run/sum> <next largest> <count> <count/sum> <running> <run/sum> <next largest> <count> <count/sum> <running> <run/sum> ... TOTAL <sum> 100% <sum> 100% Read the cut where Cumulative % first crosses 75% to 80%. Everything above that line is the vital few.
Two rules keep a Pareto chart honest. Count the same unit all the way down the column, so do not mix incidents with hours. And measure over a period long enough that one unusual week cannot dominate the ranking, which in most work settings means a quarter rather than a fortnight.
Five ways the 80/20 rule gets misused
The Pareto principle is misused more often than it is applied, and nearly always in the same five ways. Each one is easy to spot once you know the shape of it.
- Treating 80 and 20 as fixed. Four of the five examples above are not 80/20, and one is not remotely close. Fix: report the pair you actually measured and let it be 73/20 or 63/1.
- Expecting the numbers to add to 100. The two figures count different things. Fix: say “the top 20% of buyers produced 73% of sales” in full, which makes the error impossible to commit.
- Applying it forward as a prediction. A concentration measured last quarter is a description of last quarter. Fix: re-run the analysis rather than assuming the curve held, especially after you have acted on it.
- Dropping the tail. The other 80% usually contains payroll, compliance, and replying to people. Fix: use the analysis to decide where additional effort goes, not which obligations disappear. Juran renamed it the useful many for this exact reason.
- Skipping the measurement. “80% of our value comes from 20% of our work” is a hypothesis, and most people never test it. Fix: build the five-column table above before quoting a ratio in a meeting.
Finding the vital few in Taskly
A Pareto analysis is only as good as the record it runs on, which is where most teams come unstuck: the work happened, but nothing captured it in a form you can sort. In Taskly the raw material is already there, because every task carries an owner, a project and a completion date, so a quarter of finished work can be grouped and ranked without reconstructing anything from memory.
Otto, our AI assistant, is useful at the messy end of this. Ask Otto in plain English to group last quarter’s completed tasks by project and it drafts the grouping for you, as a proposal you review and approve before anything changes. Otto acts with your consent rather than reorganising your board and telling you afterwards, which matters most on exactly this kind of analysis, where the grouping decisions are judgement calls.
If you are weighing up tools for this kind of work, our breakdown of task management software for teams is a reasonable next stop, and what deliverables are in project management covers how to define the outcomes a Pareto analysis needs you to count in the first place.
