The 80/20 Rules
One of the reasons I write this newsletter is a piece of advice printed on bags of Purely Elizabeth granola:
"Follow the 80/20 rule. 80% of the time be your healthiest self. 20% of the time, indulge – guilt-free."
I've been fascinated by this statement ever since the first time I saw it.
For one thing, statistics like this rarely appear in pre-20th century marketing. You may see a lot of text on product labels (see below), but stats are almost never included.
The idea that statistical data could be informative illustrates one of the ways our society has become "statisfied", or transformed by the widespread use of statistics. It's not any more surprising to see the 80/20 rule on a bag of granola than it is to see, on the home page of the Washington Post this Tuesday, for example, statistics on declining rates of COVID-19, on election results in Brazil, on greenwashing among utility companies, and on scientific comparisons of the health benefits of coffee vs. tea (spoiler alert: coffee wins, but tea is still good for you).
Purely Elizabeth's 80/20 rule also intrigues me because it's so specific. Why not just encourage healthy behavior "most" of the time? Perhaps the answer is that being "statisfied" causes us to find specific numbers more persuasive than vague terms like "most." But specificity comes with a cost. If Purely Elizabeth had merely advised me to be healthy "most" of the time, I would've nodded and continued eating my granola without further thought. Instead, the first time I saw the 80/20 rule, I immediately wondered: Why not 70/30? Would 76/24 be ok? How could the numbers be so specific? Even if we trust the 80/20 statistic, what does it mean anyway to "indulge"? Is the company suggesting that it's alright to do cocaine two days in a row, so long as you abstain for the next eight? Somehow I doubt it.
What I find most interesting about Purely Elizabeth's 80/20 rule is that when you look back at where it came from, you find lots of other 80/20 "rules" varying in plausibility and generality, and in the ways they're understood (or misunderstood) by the general public. Taken together, these rules form a sort of museum of some of the ways our society has been "statisfied."
The 80/20 Diet
The source of Purely Elizabeth's 80/20 rule is a 2012 book called "The 80/20 Diet", written by Teresa Cutter, an Australian chef and creator of nutritional supplements. Cutter's advice is this:
"Eat well 80 per cent of the time, and 20 per cent of the time you can enjoy a little indulgence. So, six days a week, try eating foods that are good for the body.....Make one day a week your 20 per cent."
There's so much wrong – and right – with this advice.
Indulging one day per week, as Cutter recommends, would actually be 14.28% of the week (1 divided by 7), so what's really being proposed here is an 86/14 Diet. (A minor issue with the math rather than the diet.)
The main problem with Cutter's advice is that there's no research supporting a specific 80/20 ratio. I don't think there ever could be. "Indulgence" is highly subjective. For some people, it could mean a glass of wine with dinner. For others, it's all-day foraging at the Texas State Fair (they're offering deep-fried honey this year). Even identifying foods that "are good for the body" becomes challenging when you factor in method of preparation and portion control. One egg a day is good for your body; six would be devestatingly too much cholesterol.
On the other hand, research does suggest that people should eat well most of the time. And Cutter does caution that one should indulge in moderation rather than binging. So, if you don't take the 80/20 statistic too literally, she seems to be offering reasonable advice.
In sum, from a statistical perspective the 80/20 Diet is too vague and too specific at the same time. It's too vague in the sense that "indulgence" can't be clearly defined. It's overly specific in the sense that there's no evidence that 80/20 (or 86/14) is an optimal ratio. This is an example of false precision. We have no way of knowing whether 80/20 is better than 75/25, or 71/29, etc., though it would make sense to say that eating well "most of the time" is good for us. In the end, I believe Cutter chose the name "80/20 Diet" because the statistics provided a veneer of credibility. There's also a reason she latched onto this particular ratio.
The Pareto principle
Ultimately, the name of the 80/20 Diet can be traced back to the Pareto principle, also known as the 80/20 rule. This rule was named in honor of economist Vilifred Pareto, who observed in 1906 that roughly 80% of the land in Italy was owned by about 20% of the population.
Pareto's finding was later dubbed "the 80/20 rule" and popularized by a famous management consultant, Joseph Juran, who identified numerous applications in business settings – 80% of sales are generated by 20% of clients, 80% of productivity comes from 20% of workers, 80% of quality problems can be traced to 20% of causes, etc. But Juran was wise enough not to depict the rule too literally. What he stressed is not that any of these numbers are exactly 80% and 20%, but rather that the percentage for outcomes in each case is much higher than the one for causes. A lot of sales are generated by a relatively small percentage of clients, for instance. A lot of quality problems arise from relatively few sources. Etc. From a management perspective, thinking about outcomes and causes in this way can be useful, even without the specificity of an exact 80/20 ratio.
Now you can see that Purely Elizabeth (which makes excellent granola, by the way) is either Purely Deluded or Purely Deceptive in recommending 80% healthy eating and 20% indulgence. Same goes for Teresa Cutter's 80/20 Diet. No studies point specifically to this ratio as ideal for dietary health, and it derives no legitimacy from the Pareto principle or Juran's applications of it. In fact, not only were Pareto and Juran not making observations about diet, they weren't even presenting the same kind of statistic. In the 80/20 Diet, the numbers have to add up to 100%. If Cutter had recommended healthy eating 75% of the time, she'd recommend 25% indulgence. However, the Pareto/Juran numbers are independent of each other and don't have to sum to 100%. Pareto might've found that 20% of the Italian population owns 50% of the property, for example.
In the remainder of this newsletter, I'll focus mainly on 80/20 rules involving separate variables, where the numbers don't have to add to 100%.
What are 80/20 rules?
The 80/20 rules that Juran described, as well as others outside of the realm of business, can ve viewed in one of two ways.
1. The 80/20 rules reveal something important about the world. The fact that 80% of software errors and crashes are caused by 20% of bugs, for example, shares a deep connection with the fact that 80% of productivity is generated by 20% of workers. The observed ratios may deviate slightly from 80% and 20%, but not much.
2. The 80/20 rules reflect unrelated phenomena, and the true ratios may differ substantially from 80/20. At most, these "rules" only illustrate that in some settings, the majority of outcomes are produced by a small subset of causes.
I really want to believe interpretation #1, because it's both informative and fun to discover mathematical patterns underlying familiar phenomena. One of the most famous examples of this is the normal distribution (aka the bell curve). If we allow for a little definitional flexibility, we find lots of normal distributions in the world – e.g., blood pressure, IQ, the wingspans of birds, and even the retirement ages of NFL players (see below).
Sadly, it appears that #2 is the better interpretation.
—In some cases (e.g., Purely Elizabeth and Teresa Cutter), there's no evidence for an 80/20 rule.
—In some cases, the 80/20 ratio is observed in a study, but nobody (including those who conducted that study) would expect that ratio to generalize to the point that we could call it a "rule". For example, even though the 80/20 ratio has been estimated for productivity in some companies, the percentage of workers responsible for 80% of productivity varies widely across industries and comapnies depending on the nature of the work, the size and culture of the company, and so on. Likewise, the notion that 20% of bugs cause 80% of software errors and crashes is an observation that Microsoft made in 2002, in one analysis, and doesn't generalize over time.
—In some cases, researchers identify an 80/20 ratio in their study and portray it as a general rule, but later research proves them wrong. For example, an influential 1997 study claimed that when we look at the transmission of infectious diseases, 20% of infected individuals tend to be responsible for 80% of the transmission. Later research showed that when super-spreading occurs, the numbers don't generally correspond to an 80/20 pattern. Rather, all we can say is that at certain times during the spread of an infectious disease, relatively small proportions of infected individuals account for disproportionate amounts of transmission.
—In some cases, when an 80/20 ratio is observed in a study, the researchers don't claim that it's a general rule, but others do, and the misconceptions take root. Consider, for example, the contention that 20% of the world's population controls 80% of wealth. This statistic, commonly cited in discussions of income inequity, comes from a 1992 United Nations Development Programme (UNDP) report. It's sometimes called the "champagne glass effect" owing to the way it has been depicted visually (see below).
Income inequity continues to be extreme, but the ratios vary depending on the time period considered, the scope of the analysis (the entire world vs. individual countries), and the operational definition of affluence. For example, in a 2021 report, UNDP partners showed that 76% of the world's wealth is controlled by roughly 10% of the population, with considerable variation from country to country. The global "rule" now is roughly 80/12, not 80/20. Continuing to invoke the 80/20 rule in discussions of economic inequity is inaccurate and misrepresents the fact that, globally, the problem is getting worse.
—In a few cases, the 80/20 ratio seems genuine, in the sense that something close to that ratio may be generally true for individual phenomena. For example, an influential 2010 study showed that elite athletes tend to devote roughly 80% of their training to low-intensity activities, with the remaining 20% focused on high-intensity ones. However, (a) experts continue to question this particular finding, and (b) even if credible, it has nothing to do with any other 80/20 rule one might find (like 20% of Italians owning 80% of the country's land). It's a coincidence that the ratio is 80/20 in each case. If you looked long enough, you could find other coincidences (e.g., 75/25 "rules").
Public perceptions
Poking around the internet reveals a lot of confusion about 80/20 rules. There are claims that the 80/20 Diet reflects scientifically-determined percentages, for example, along with claims that 20% of employees are generally responsible for 80% of outcomes such as productivity and complaints.
You can also find confusion – or intentional deception – among those who make money from books with "80/20" in their titles. For example, in a 1999 best-seller called "The 80/20 principle: The secret to achieving more with less", you'll find the following:
"The 80/20 principle has two almost opposite appeals. On the one hand, it is a statistical observation, a proven pattern – solid, quantitative, reliable, hard.... On the other hand, the principle has a totally different side – soft, mystical, eerie, almost magic in the way that the same pattern of numbers crops up everywhere... The sense that we are connected to each other and to the universe by a mysterious law...generates a sense of wonder and awe."
Here I'm experiencing more of a sense of bulls***t.
This is a revealing passage. The author seems to imply that if something is a "statistical observation", it must be accurate and objective. As you know, statistics vary in accuracy, generalizability, and meaningfulness. The author's reference to the 80/20 rule cropping up "everywhere" also illustrates confirmation bias. Of course you can find 80/20 examples everywhere. You can find lots of 62/38 examples too. And 75/25 examples. There's no evidence that 80/20 rules are especially prevalent. For example, here are some 70/30 "rules" that have been proposed:
—70% of weight loss is attributable to the foods you eat; 30% is attributable to exercise. (This is not generally true. The ratio varies widely depending on your health, your weight, your eating habits, your exercise routines, etc.)
—70% of students fail to perform at grade level; only 30% of students actually do. (This is not generally true. The ratio varies widely depending on school district, school, student grade levels, subject, etc.)
There are also 70/30 rules that are reminiscent of the 80/20 Diet, in the sense that they're just recommendations with no specific evidence to support them. For example:
—Spend 70% of your earnings each month; add the remaining 30% to your savings.
—To obtain support for a project, spend 70% of the time describing details of the project and 30% focusing on establishing trust.
—As a salesperson, allow your clients to talk 70% of the time; only do 30% of the talking yourself.
—Strive to be great 70% of the time; allow yourself to be mistaken 30% of the time.
I'm not saying that any of this is bad advice. All of it may be wise and useful. I'm only pointing out that there's nothing special about 70/30 rules, or 80/20 rules, or any of the others. What these rules tell us is that we've been "statisfied – we're accustomed to seeing ideas framed in statistical terms, and the presence of statistics may increase credibility and persuasiveness.
Wikipedia
Unfortunately, Wikipedia, one of the most influential sources of information on the planet, is a potential source of confusion on this topic. The "Pareto principle" wiki is not only badly written (and flagged as having multiple issues), some of the 80/20 rules it cites are either incoherently described or were never intended to be 80/20 rules in the first place. In short, the Wikipedia page reflects the worst kind of confirmation bias, where anything that looks like an 80-20 split is glommed together, accompanied by unsubstantiated hints of a universal principle at work.
Conclusion
In the early 20th century, after a period of relatively slow development, statistical methods and data began to proliferate with increasing speed. One consequence of these changes is that society became "statisfied", meaning, among other things, that for the first time in history, the general public became accustomed to seeing statistics presented as a form of evidence. Statistics became one more way of supporting arguments and selling products. One of the goals of this newsletter, as well as the website and book I'm developing, is to describe how we've been transformed by these changes.
There's nothing inherently bad about the "statisfying" of society. Statistics are highly informative, sometimes in ways that nothing else can approximate. But statistics can also be misused and misinterpreted. I think the 80/20 rules are a good illustration of both. Looking back at how these rules have been presented and discussed, we find the following:
—In some cases, 80/20 rules are presented without evidence, because statistics add a veneer of legitimacy or credibility.
—In some cases, 80/20 rules are presented on the basis of evidence that doesn't justify such specificity. Although it's useful to know when a majority of outcomes stem from a minority of causes, statistical definitions of "majority" and "minority" reflect false precision.
—In some cases, 80/20 rules are presented on the basis of studies that weren't designed to identify rules. These studies report specific ratios that don't generalize (e.g., because the data are time- or place-specific). The fact that the results don't generalize isn't reflective of methodological limitations.
—In a few cases, 80/20 rules are presented as some sort of fundamental pattern in nature. This idea is untenable.
—In a few cases, 80/20 ratios are presented as general rules, but only with respect to specific phenomena (e.g., training routines among elite athletes). If these findings are supported by further research, they'll be the only genuine examples of 80/20 rules we have.
Thanks for reading. Hopefully more than 20% of this newsletter captured more than 80% of your attention!