Average Means Failure: Why Products Designed for Everyone Sell to No One

In 1945, a physician named Robert Latou Dickinson and a sculptor named Abram Belskie unveiled a pair of white alabaster statues at the American Museum of Natural History in New York. The female figure was called Norma. She was built from the averaged measurements of fifteen thousand young American women, every dimension calculated to the fraction of an inch: height, weight, bust, waist, hips, thigh, calf, ankle, foot. She was, by the math, the statistically average American woman.

Time magazine ran a feature. CBS broadcast a documentary segment where Norma's measurements were read aloud so women at home could check themselves against the standard. The Cleveland Health Museum acquired the statues and the Cleveland Plain Dealer organized a contest: find Norma in the flesh. One hundred dollars in war bonds to the woman whose body matched the statue on all nine dimensions.

Nearly four thousand women submitted their measurements.

Not one matched.

Fewer than forty were average on even five of the nine dimensions. The winner, a twenty-three-year-old theater cashier named Martha Skidmore, matched on roughly five. She was the closest anyone came to the "average" woman, and she wasn't particularly close. The statistical average of fifteen thousand women described not even a single real person.

The contest organizers treated this as a curiosity. A charming surprise for the morning paper. They missed the deeper problem entirely, because the flaw wasn't in the contest. The flaw was in the concept. And it had been hiding in plain sight for over a century, since a Belgian poet looked up at the stars and made a mistake that would shape how we design nearly everything.

The Astronomer's Error

Adolphe Quetelet was not, originally, the kind of man you'd expect to flatten humanity into a single number. Born in Ghent in 1796, the fifth of nine children, he lost his father at seven and spent his youth painting, writing poetry, and co-writing an opera with his friend Germinal Dandelin. He published verse that critics praised. He translated tales of chivalry from four languages into French. He was, by every account, the kind of person who noticed the differences between things.

Then, in 1823, the Belgian government sent him to Paris to study astronomy. There he met Pierre-Simon Laplace and Joseph Fourier, two of the greatest mathematicians alive, and he learned the tool that would undo his appreciation for variation: the normal distribution.

In astronomy, the normal distribution solves an important problem. When you measure the position of a star multiple times, each measurement is slightly different. Not because the star moved, but because your instruments are imperfect. The variation is noise. The average of all your measurements gives you the best estimate of the star's true position. This is elegant mathematics for a simple situation: one fixed object, many imprecise readings.

Quetelet returned to Brussels and did something no one had done before. He took the method designed for measuring stars and applied it to human beings. When he analyzed the chest measurements of 5,738 Scottish militiamen, he found they formed a bell curve. And he interpreted that curve the same way an astronomer would interpret readings of a distant star: the average was the signal, and every real soldier was noise.

He published his findings in 1835 in a book called Sur l'homme (On Man). In it, he introduced a concept that would outlive him by two centuries: l'homme moyen. The average man. Not a statistical convenience, but an ideal. Quetelet wrote that if the average man were fully determined, he might be considered "the type of perfection," and that anything deviating from his proportions would constitute "deformity or disease... monstrosity."

The Athenaeum, the most prestigious literary publication in the English-speaking world at the time, called the book "an epoch in the literary history of civilization." Florence Nightingale believed Quetelet's statistical patterns were God's laws made visible. Karl Marx drew on his framework in developing the concept of socially necessary labor time. Prince Albert studied under Quetelet. The idea spread through science, government, and industry like a contagion with no immune response.

And it was wrong. Not wrong in the way that early theories are sometimes imprecise. Wrong in a structural way. A star has a true position. A human being does not have a "true" body from which they deviate. A soldier with a 36-inch chest and a soldier with a 44-inch chest are not errors in the production of a 40-inch man. They are different men. Quetelet had confused diversity with malfunction. As Harvard professor Todd Rose put it nearly two centuries later: "The average became normal and the individual became error."

The Ghost in the Cockpit

The military was among the first to run Quetelet's framework at scale. During the Civil War, the Union Army averaged the measurements of hundreds of thousands of soldiers and used the results to design standardized rations, weapons, and uniforms. Before that, military clothing had been custom-sewn. After Quetelet, soldiers were sorted into small, medium, and large. Those categories eventually migrated to civilian clothing, and then to everything else.

In 1926, the Army took the next logical step. It measured several thousand male pilots, averaged their dimensions, and designed airplane cockpits to fit the result. The seat height, the distance to the pedals, the position of the control stick, the shape of the helmet. All built for one person: the average pilot.

By the late 1940s, pilots were dying at alarming rates, and no one could figure out why. The planes had improved. The training had improved. The crash rate hadn't. When the Air Force finally looked at the cockpit itself, they called in a twenty-three-year-old named Gilbert Daniels.

Daniels was a physical anthropology fresh from Harvard. His senior thesis had examined the hand shapes of hundreds of male students, all from similar backgrounds, all roughly the same age. Even in that homogeneous group, when he averaged the measurements, the "average hand" didn't resemble any individual's actual hand. The finding had stayed with him.

At Wright Air Force Base, Daniels and his colleagues measured 4,063 active-duty pilots across 140 physical dimensions. He selected ten that mattered most for cockpit design: height, chest circumference, arm length, hip breadth, sitting height, and five others. Then he checked how many pilots fell within the average range on each.

On three of the ten dimensions, fewer than 3.5 percent qualified as average.

Out of 4,063 pilots, not a single one was average on all ten dimensions.

Daniels published his findings in a 1952 technical note with a title that contained a question mark: The "Average Man"? His conclusion was quiet and devastating. The tendency to think in terms of the average man, he wrote, was "a pitfall into which many persons blunder." It was "virtually impossible to find an average airman not because of any unique traits in this group but because of the great variability of bodily dimensions which is characteristic of all men."

The cockpits weren't designed for anybody. They were designed for a ghost — a statistical composite who didn't exist in the population he was derived from. And pilots were crashing because their seats didn't adjust, they couldn't reach their pedals, their helmets didn't fit.

The Air Force mandated adjustable cockpits. Adjustable seats. Adjustable pedals. Adjustable helmet straps. They stopped designing for a fictional person and started designing for real variability. The accident rate dropped from 23.6 aircraft destroyed per 100,000 flying hours to 4.3.

More than a fivefold reduction. Because someone stopped pretending the average was real.

What Happens When You Average a Brand to Death

In 1972, three friends from New York started a juice company. Leonard Marsh and Hyman Golden ran a window-washing business. Arnold Greenberg owned a health food store in the East Village. They called the company Unadulterated Food Products, started supplying fruit juices to health food stores in their spare time, and went about it so casually that all three kept their day jobs.

The breakthrough came in the late 1980s, when they figured out how to bottle tea while hot, eliminating preservatives. By 1991, the company, which they'd renamed Snapple, was doing a hundred million dollars in revenue. By 1994, it was six hundred and seventy-four million. Sales had roughly doubled every year for three consecutive years.

Snapple was specific in a way that almost no large brand manages to be. The distribution ran through independent operators serving bodegas, delis, gas stations, and lunch counters, what the industry calls the cold channel. In a bodega cooler, Snapple wasn't competing against a hundred other options on a warehouse shelf. It was one of a handful of interesting things, and the quirky label and unusual flavors grabbed attention. You grabbed one because it caught your eye. The marketing was equally specific: Wendy Kaufman, a real employee from the shipping department, read actual fan mail on national television. Howard Stern was paid to make fun of the product on air. Rush Limbaugh had started recommending it for free before anyone thought to make it a deal. These were relationships with polarizing, personality-driven hosts whose audiences were intensely loyal. The brand felt like it belonged to the people who drank it.

Then, in November 1994, Quaker Oats bought Snapple for $1.7 billion.

William Smithburg, Quaker's CEO, had reason to be confident. Quaker had transformed Gatorade into a powerhouse using disciplined mass-market strategy: supermarket distribution, athletic sponsorships, streamlined product lines. He told the press that Quaker's team knew how to "advance Snapple as well as Gatorade to the next level." Wall Street immediately disagreed, estimating Quaker had paid roughly a billion dollars too much.

What happened next was systematic. Quaker's beverage president, Donald Uzzi, offered Snapple's independent distributors a deal: take on Gatorade's distribution in exchange for giving up part of their exclusive Snapple territories to Gatorade's existing system. It was an attempt to merge a cold-channel brand with warm-channel logistics, to move a product people grabbed from a bodega cooler into a supermarket aisle where it would sit at room temperature next to Coke, Pepsi, and a hundred other options. Quaker fired Wendy Kaufman and replaced her authentic, unpolished commercials with corporate campaigns produced by Gatorade's advertising agency. They dropped Howard Stern in April 1995, claiming he'd become "overly commercialized." They dropped Rush Limbaugh because he was "too controversial." They cut the lineup of quirky flavors. They redesigned the packaging. They applied the Gatorade playbook to a brand that ran on a completely different set of rules.

John Deighton, writing in the Harvard Business Review, later identified the core problem as "a fatal mismatch between brand challenge and managerial temperament." Snapple's brand vitality, he wrote, "responded better to play than to planning."

Twenty-seven months after paying $1.7 billion, Quaker sold Snapple to Triarc Companies for $300 million. A loss of $1.4 billion in just over two years. $1.6 million for every day they owned it.

Triarc's new head of Snapple, Mike Weinstein, reversed every decision Quaker had made. Brought back Wendy. Brought back Stern and Limbaugh. Restored the independent distributors. Brought back the quirky flavors and the original marketing tone. His operating philosophy was a single sentence that said everything Quaker had failed to understand: "We're not in the soft-drink business; we're in the fashion business."

Three years later, Triarc sold Snapple to Cadbury for $1.45 billion.

The math shows the impact being average has on a product's value. Quaker took a specific brand, made it average, and destroyed $1.4 billion. Triarc took the wreckage, made it specific again, and, in roughly the same amount of time, created $1.15 billion in value. The product hadn't changed. The tea was the same tea. What changed was whether the brand was allowed to be something specific, something that some people loved and most people had no opinion about, or whether it was forced into being something generic, something inoffensive, something designed to appeal to the broadest possible audience.

Average.

Average Is Invisible

The reason this happens is based in neurology, and it's simpler than you'd expect.

In 1933, a German psychologist named Hedwig von Restorff ran a series of memory experiments. She showed subjects lists of items, mostly similar, with one distinctly different. A list of ten numbers with a single word in the middle. A series of blue shapes with one red one. The finding was consistent and robust: people remembered the item that was different. Of course they did. They hadn't tried harder to memorize it, and it wasn't objectively more important. The brain is simply wired to notice contrast and discard sameness.

This is now called the Von Restorff effect, or the isolation effect, and it's survived nearly a century of replication. The brain is a difference-detection machine. It doesn't process every piece of information equally. It filters. And its primary filter is: does this stand out from the pattern, or does it match?

If it matches, it gets less processing. Less attention, less memory encoding, a weaker emotional response. Not zero — but close enough. And in a competitive market where a customer encounters dozens or hundreds of options, "less" is functionally equivalent to invisible.

Think about the last time you walked down a grocery aisle and actually noticed a product. Not scanned past — noticed. Something about it broke the pattern. A shape that didn't match, a name that surprised you, a color that contrasted with the shelf. Your brain flagged it because it violated the expectation your visual cortex had already built from the surrounding products. That violation is the signal. Everything that matched the pattern was wallpaper.

Average products don't get average attention. They get filtered. They become invisible.

The Dumbest Possible Name

In 2009, Mike Cessario was standing at the Vans Warped Tour watching bands play in the summer heat. Cessario had grown up in the punk and hardcore scene outside Philadelphia, played guitar in multiple bands, designed flyers and album covers, and eventually channeled his visual instincts into a career in advertising. He'd spent the better part of a decade as a creative director at agencies like VaynerMedia and Crispin Porter, working on campaigns for Netflix, Toyota, and Mountain Dew. He knew how branding worked. He also knew what was bothering him.

The musicians on stage were pouring water into Monster Energy cans. Monster was a sponsor. The cans looked cool on stage: tall, dark, aggressive. Water bottles didn't. So the performers dumped the energy drink and refilled the can with water, because holding the right container mattered more than drinking the right liquid. Cessario watched this happen and a question crystallized: why does all the fun, irreverent, memorable branding belong to junk food?

He sat on it for eight years.

In 2017, he trademarked the name Liquid Death. In 2018, still without a product, he spent $1,500 on a fake commercial, put it on Facebook, and promoted it with a few thousand dollars in paid media. Within four months, the video had three million views and the page had eighty thousand followers. More than Aquafina. For a can of water that didn't exist yet.

"Nobody was going to just write a check for that idea, because it was so out there," Cessario said. The name was wrong. The imagery — skulls, gothic typography, heavy metal aesthetic — was wrong. The tagline, "Murder Your Thirst," was wrong. Every choice violated every convention of the $350-billion bottled water industry, where the rules had been settled for decades: blue packaging, mountain imagery, words like "pure" and "natural" and "clean." The entire category looked the same. That was the point.

"You kind of have to trick your brain to come up with a bad idea to truly be thinking in innovative territory," Cessario told CNBC. "What's the dumbest possible name for a super healthy, safest beverage possible? Liquid Death. Probably the dumbest name."

Science Inc. wrote the first check. The first year brought three million in revenue. When the cans hit Whole Foods, they became the fastest-selling water brand on the shelf. Then came the marketing campaigns that made the polarization strategy explicit: a ten-song death metal album called Greatest Hates, with lyrics pulled verbatim from actual hate comments about the brand. Track titles included "Dumbest Name Ever for Water" and "This Crap Is Pure Evil." A hundred limited-edition skateboards painted with Tony Hawk's actual blood, priced at five hundred dollars each. They sold out in twenty minutes.

By 2024, revenue had reached $333 million. A hundred-and-ten-times growth in five years. Available in over 130,000 retail locations. Valued at $1.4 billion. In a category where every brand had spent decades trying to look trustworthy and pure, the company that put skulls on a water can and told people to murder their thirst built a cult following that existing brands couldn't touch.

Cessario described the business as "an entertainment company that monetizes through beverage." He told the press that his marketing team operated more like Saturday Night Live. The hate was the strategy. Anyone who thought the brand was stupid was never going to buy water based on a mountain on the label anyway. They weren't the customer. And by not being the customer, they made the brand more intensely meaningful to the people who were.

The Shirt

If you were to design a shirt for one person, you could make it perfect. The exact measurements, the fabric they prefer against their skin, the color that makes them feel like themselves, the cut that fits the way they actually move. A shirt like that costs at least five hundred dollars because it takes a tailor who listens, a pattern that adjusts, and a refusal to compromise on any dimension. The person who wears it knows immediately that it was made for them. Not for someone like them. For them.

Now try to design one shirt for eight thousand people. You can't ask what color they prefer, because they'll give you eight thousand answers. You can't optimize the cut, because their bodies are all different in ways that won't average into a useful shape. The Norma contest proved that. The pilot study proved that. You end up with a shirt that is nobody's favorite color, doesn't quite fit anyone's body, and is made from a fabric chosen for cost rather than feel. It retails for fifty dollars. It sits on a rack next to a hundred other shirts that made the same compromises. Maybe somebody buys it.

No tailor works by averaging their clients' measurements. You'd get a garment that fits the room but not a single person in it.

This is the logic of the average applied to products. This is what average does to the value of a product, and fashion brands lose tens of billions of dollars annually as a result. Standard sizing was built on Quetelet's framework, the same bell-curve thinking that produced Norma, the same assumption that there's a "true" body that real bodies deviate from. There isn't. There never was.

And it isn't only clothing. Pharmaceutical companies design drugs for the "average patient" and find that between fifty and seventy-five percent of those drugs don't work as intended. Humira, one of the best-selling drugs in history, produces remission in roughly one in four rheumatoid arthritis patients who take it as a standalone treatment. The drug isn't bad. It's the right medication for some, just not most. There is simply no average patient, just as there was no average pilot and no average American woman. In 1984, Benjamin Bloom compared students who received one-on-one tutoring with students taught in traditional classrooms. The tutored students outperformed the traditional classroom students by two standard deviations. The bespoke version of education doesn't just beat the average version. It demolishes it.

The pattern repeats in every domain where anyone has bothered to check. Standardized instruction produces mediocre learning. One-dose-fits-all medicine produces unpredictable responses. Off-the-rack clothing produces returns. And cockpits built for the average pilot produce crashes. Wherever someone has designed for the composite instead of the person, the result is the same.

The Winner-Take-All Problem

For entrepreneurs, average doesn't just underperform. It produces zero.

When a customer stands in front of a shelf, they don't buy the second-best option for their needs. They buy the best one. If two products are roughly equivalent, if neither stands out, if both are competent but unremarkable, the brain struggles to generate a strong enough signal to choose. This is the jam study phenomenon. Researchers Sheena Iyengar and Mark Lepper found that shoppers were ten times more likely to purchase when choosing among six distinctive jams than among twenty-four. Fewer options aren't always better. But distinct options produce clean signals, and similar options produce noise.

In markets with real competition, a customer only buys the product they believe is the best option for them. Not the best product overall — the best for them. There is one winner for each customer. The runner-up, third place, fourth place, they all end up in the same position: no sale. This is a winner-take-all dynamic at the individual level, repeated across every customer in your market.

An average product cannot win this dynamic. It isn't bad. It just isn't the best option for anyone in particular. A product designed for the hypothetical average customer is, by definition, a product optimized for no one. It's the cockpit that fits no pilot, the shirt that flatters no body, the Snapple that tastes like every other drink on a supermarket shelf. It doesn't trigger the Von Restorff response. It doesn't generate the signal that resolves into a purchase decision. It sits there, competent yet forgettable, while the customer picks the thing that feels like it was made for them.

You would think the safe play is to build something that offends no one. Something broadly acceptable. Something that, if you tested it with a hundred people, most of them would say it's fine. And a hundred people saying "it's fine" feels like validation. It feels like a big market. It feels safe.

It's not. It's the Norma contest. Nearly four thousand women, zero matches. One hundred polite nods, zero purchases.

It is better to build something that ninety percent of people actively dislike and ten percent cannot live without than to build something that everyone finds acceptable and no one finds essential. The ten percent will buy. They'll tell their friends. They'll come back. The ninety percent? They were never going to buy from you anyway. They'll buy from whoever built the thing that feels like it was made for them. The only losing move is to build for the middle, the composite, the average. Because the middle is where the ghost lives. In the real world average doesn't exist. It's a statistical fiction that describes no one and sells to no one.

You might wonder about the companies that seem to contradict this. Amazon, Walmart, Coca-Cola. Don't they serve everyone? They do now. They didn't start that way. Amazon sold books. Only books. Jeff Bezos chose the single most catalogue-friendly product category, sold it on a platform that was niche at the time, and optimized everything for that one use case before he touched anything else. Facebook launched at a single college dormitory and didn't open to the general public for over two years. Tesla built one car, a six-figure electric sports car for wealthy early adopters who wanted to make a statement about the future. Every company that looks like it serves everyone has an origin story built on radical specificity. They started by being the best option for a very small number of people, and they expanded from that foundation. You can start narrow and go broad. You cannot start broad and go narrow. Broad means average, and average means you never tap into the psychological desire needed to win over customers.

Quaker tried to take Snapple broad, and the brand evaporated. They lost a billion dollars. Triarc made it narrow again, and they made that billion dollars back.

Specificity scales. Averages don't.

Adolphe Quetelet was a poet before he was a statistician. He painted and wrote verse and co-wrote an opera. He was, by every account, a man who noticed the particulars of things. Then he went to Paris and learned a technique for measuring stars, and he came back and applied it to people, and in doing so he created a framework that has shaped two centuries of how we design, build, market, and sell. At first glance, the framework is elegant. But it is wrong. Not wrong in the details but wrong in the premise: that there is a true version of a person, that variation from that version is error, that the average is the signal and the individual is the noise.

The individual is the signal.

Norma was a ghost. The average pilot was a ghost. The average patient, the average student, the average customer. All ghosts. Statistical averages that describe the room but fit no one in it.

If you build for the ghost, you will build something that sits on the shelf, competent and unremarkable. Your customer will walk right past it toward the thing that feels like it was made for them. The brand with the skulls. The drink from the bodega. The shirt that fits.

Build for one person. The right person. Make something so specific that most people won't want it and one person won't be able to imagine life without it. Because in every market, in every category, across every domain where anyone has measured it, the same result comes back: average products don't get average results. They get zero.

You must build for one, or you will build for none.

---

The idea that average is safe is one of the most expensive misconceptions in business, and it starts earlier than most people realize. Chapter 7 of Ideas That Spread traces the full framework behind Build for One or Build for None, the strategy of finding the single person whose problem is so urgent that they'd duct-tape together a solution if you didn't show up, and building from there. The chapter goes deeper into why the smallest viable market isn't a constraint but a competitive advantage, and how the companies that look like they're for everyone all started as products for someone. The blog showed you why average fails. The book shows you how to find the one.