The Product That Fit 4,000 People and Nobody at All

In 1950, the United States Air Force had a problem it couldn't explain. Pilots were losing control of their aircraft at alarming rates. The jets were faster and more capable than anything from the previous generation, but the crash rate had spiked well beyond what new technology could account for. The military investigated mechanical failure. It investigated training. It investigated pilot error. Nothing fit.

Then someone asked a different question: what if the cockpit itself was wrong?

The Air Force had designed every cockpit around the average pilot — a composite figure derived from a 1926 measurement study. Seat width, pedal distance, helmet size, control-stick placement — all built for a body that represented the mathematical midpoint of thousands of airmen measured decades earlier. It seemed reasonable. Design for the center and you serve the most people. That was the assumption, and nobody had tested it. The principle extends far beyond cockpits: products designed for "everyone" fit nobody, and the fastest path to scale is building for one specific person first — then expanding.

A twenty-three-year-old researcher named Gilbert Daniels decided to test it.

The Myth of Average

Daniels wasn't an engineer. He was a physical anthropologist who'd studied the variability of hand shapes at Harvard, where he'd noticed something that bothered him: the "average" hand in his dataset didn't actually match any real hand in his sample. When he joined the Air Force's Aeromedical Research Laboratory, he carried that suspicion with him.

He measured 4,063 active-duty pilots across ten physical dimensions believed to be most critical for cockpit design — height, chest circumference, sleeve length, sitting height, hip breadth, and five others. Then he defined "average" generously: anyone who fell within the middle 30 percent of the range on a given dimension counted. You didn't need to be exactly average. You just had to be close.

Out of 4,063 pilots, the number who fell within the average range on all ten dimensions was zero.

Not a small number. Not a surprising minority. Zero. The cockpit built for the average pilot fit literally nobody who flew the planes.

Daniels went further. He tested what happened if you dropped the requirement to just three of the ten dimensions. Pick any three — say, neck circumference, thigh circumference, and waist circumference. Even with that drastically reduced bar, fewer than 3.5 percent of pilots qualified as average. The "typical" pilot was a statistical ghost. He existed in the spreadsheet and nowhere else.

Daniels published his findings in a 1952 technical note titled "The 'Average Man'?" — question mark included. The Air Force's response changed the history of design. Instead of trying to find pilots who matched the cockpit, they redesigned the cockpit to match the pilot. Adjustable seats. Adjustable pedals. Adjustable helmet straps. They stopped designing for a fictional person and started designing for real variability.

The crash rate dropped.

The Niche of One

There's a sentence that circulates in startup circles like a proverb nobody follows: "If your ideal customer is everybody, your actual customer is nobody." Founders hear it, nod, and then build for the widest possible market anyway.

Gilbert Daniels proved why that instinct is wrong, more than fifty years before the first Y Combinator batch. When you design for the average — the composite, the generalized, the "broad appeal" — you design for a person who doesn't exist. The product that tries to serve everyone ends up with the exact same problem as a 1950s cockpit: it technically accommodates the middle and actually fits nobody well enough to matter.

The counterintuitive path to scale is extreme specificity. Build for one person first. Make something that fits them so well it feels custom. Then expand.

Call it the Niche of One.

The evidence for this is everywhere, hiding in the origin stories of companies that now seem inevitable. Tobias Lutke wanted to sell snowboards online in 2004. Every e-commerce platform he tried was clunky, bloated, and clearly built for "any business selling anything." None of them worked well for a small merchant selling a specific product to a specific buyer. So he built his own storefront from scratch using Ruby on Rails. The snowboard shop was called Snowdevil, and it did fine — but people kept asking about the platform underneath it. Over the next year, Lutke and his co-founders realized the tool they'd built for their own narrow problem was more interesting than the store. They launched Shopify in 2006. Today it powers millions of merchants and is the second-largest e-commerce platform in the United States behind Amazon.

Shopify didn't start by trying to be an e-commerce platform. It started by trying to sell snowboards. The platform was a side effect of solving one person's problem with total specificity.

Facebook launched in February 2004 exclusively at Harvard. Not colleges in general. Not Ivy League schools. Harvard. Mark Zuckerberg didn't build a social network for everyone — he built a digital version of the Harvard face book, the physical directory of student photos that every freshman class received. The initial product only made sense if you were a Harvard undergraduate. Within the first month, more than half the undergraduate student body had signed up. Then it expanded to Columbia, Stanford, and Yale. Then to all Ivy League schools. Then to other universities. Then to high school students. Then, in September 2006 — two and a half years after launch — to the general public. At every stage, the product was so well-tuned to its current users that the next adjacent group pulled themselves in.

Stewart Butterfield and his team at Tiny Speck spent three years and millions of dollars building Glitch, an online game that never found an audience. It shut down in 2012. But during development, the team had built an internal communication tool to coordinate across offices. That tool — designed for one team, solving one specific workflow problem — became Slack. It launched publicly in 2014 and by 2018 was valued at over $7 billion. The product that was built for nobody in particular (Glitch) died. The product that was built for a handful of people in a specific situation (Slack) became one of the fastest-growing enterprise tools in history.

The pattern is the same every time. The winning product isn't the one that appeals to the most people on paper. It's the one that is indispensable to a small number of people in practice.

The Hair-on-Fire Test

Geoffrey Moore formalized this in Crossing the Chasm in 1991: start with a beachhead segment so narrow that you can dominate it completely, then use that dominance as a platform to expand outward. But knowing the strategy doesn't tell you how to pick the right beachhead. Most founders choose a niche that's convenient, not one that's urgent.

There's a sharper filter. Before you ask "who might want this?" ask a different question: "who has their hair on fire?"

A hair-on-fire problem is one so painful that the person experiencing it would pay for a solution even if the solution were imperfect, incomplete, or ugly. They're not comparison shopping. They're not reading reviews. They're reaching for anything that stops the burning. This is how you overcome the 9X problem — when the pain is urgent enough, the customer's 3X overvaluation of what they already have collapses. If your earliest customers will tolerate a rough product because the problem is that urgent, you've found a real niche. If they need to be persuaded, educated, or convinced of the value, you haven't.

Airbnb's earliest users weren't travelers who thought it might be charming to stay in someone's apartment. They were people who couldn't find a hotel room during a design conference in San Francisco where every hotel was sold out. The problem was immediate, physical, and binary — you either had a place to sleep or you didn't. Brian Chesky and Joe Gebbia didn't need to explain the value proposition. They needed to provide an air mattress and breakfast.

The hair-on-fire test eliminates the most dangerous kind of product-market-fit illusion: the product that lots of people find mildly interesting and nobody finds essential. Mild interest doesn't convert to purchases. It converts to free-trial signups that churn after a week and survey responses that say "neat idea" — the same polite lies customers tell you when they say "I love it" — while the customer continues doing whatever they were doing before.

What Happens When You Design a Product for Everyone?

Ron Johnson learned this in one of the most expensive lessons in retail history.

In November 2011, Johnson became CEO of JCPenney. He'd just come from Apple, where he'd built the Apple Store into the most profitable retail operation per square foot in the country. His mandate at JCPenney was transformation. The company was stagnant. The brand felt tired. Johnson was going to fix that.

His strategy was to modernize everything at once. He eliminated the coupons, clearance sales, and promotional pricing that JCPenney was known for, replacing them with a clean "fair and square" everyday low pricing model. He redesigned the stores to include boutique-style brand shops. He dropped older merchandise lines and brought in hipper brands. He changed the logo. He was, in effect, building a new store for a new customer — one who valued design, simplicity, and transparency. Johnson had confused changing the context around the product with changing the customer it served.

The problem was that this new customer was a composite. An average. A person who existed in a strategy deck but not in a shopping mall. And the real customers — the ones who actually drove JCPenney's revenue — loved coupons. They loved the hunt for a deal. The promotional pricing wasn't a flaw to be fixed. It was the product.

In 2012, sales dropped $4.3 billion — a 25 percent decline in a single year. Fourth-quarter same-store sales fell 32 percent. The company posted a net loss of $985 million. Johnson himself later admitted he had underestimated how much the core customer depended on and enjoyed coupons. In April 2013, seventeen months after he started, the board fired him.

Johnson wasn't incompetent. His Apple Store track record proved that. But at Apple, he'd designed for a specific person — someone who already wanted Apple products and would pay a premium for the experience of buying them. At JCPenney, he designed for a hypothetical person who didn't exist in sufficient numbers, while alienating the specific people who did. He built an average cockpit.

The Expansion Paradox

The instinct to go broad early feels rational. A bigger addressable market means more potential revenue. More potential revenue means easier fundraising. Easier fundraising means faster growth. The logic is clean on a slide deck. It's wrong in reality.

Specificity creates the conditions for expansion. When you serve a narrow group exceptionally well, three things happen that broad positioning can never produce.

First, you get signal. A small, well-defined user base gives you clear feedback. You know exactly what they need, what frustrates them, and what they'd pay more for. A broad, diffuse user base gives you noise — contradictory requests from people with fundamentally different problems, making it impossible to prioritize.

Second, you get evangelism. People who feel like a product was built for them don't just use it. They tell other people who are like them — the same tribal identity dynamic that lets a tiny brand beat Nike. That's how Facebook spread from Harvard to other Ivy League campuses before anyone at Stanford had to be sold on the concept. That's how Shopify spread from Lutke's snowboard community to the broader world of small online merchants. Word-of-mouth at scale requires a community tight enough that its members actually talk to each other.

Third, you get pull instead of push. When your niche is right, expansion doesn't require a new marketing strategy. The adjacent market reaches toward you because your existing users overlap with it. Slack didn't launch a campaign targeting enterprise companies. Enterprise employees who'd used Slack at startups brought it with them when they changed jobs. The product followed the people, and the people followed their own networks.

Going broad first reverses all three dynamics. You get noise instead of signal. You get indifference instead of evangelism. You push instead of getting pulled. And you spend enormous resources trying to be adequate for everyone instead of being indispensable for someone.