It wasn’t until late 2008 that Brian Chesky had finally developed a system. He kept his credit cards in a baseball card binder, with one plastic sleeve per card, because he had accumulated so much debt that he needed a way to track it.
Chesky’s startup, an idea to let people rent air mattresses in their living rooms, had been rejected by seven investors who never took the time to write back. Eight more investors flatly refused the idea entirely. At this point, he was $20,000 in debt and burning $1,000 a week with absolutely nothing to show for it.
Chesky and his co-founders broke away from convention by doing something completely unrelated to what they’re selling. It was the 2008 election year, and the group decided to buy cheap cereal, box it in homemade packaging that read “Obama O’s” and “Cap’n McCain’s,” and sell it for $40 a piece at political conventions. This out-of-the-ordinary method worked, raising over $30,000, which was well enough to pay off debt and buy themselves a few more months. This company, with rough, unconventional beginnings, was Airbnb, and today it’s worth more than most countries' entire economies.
In frequent tellings of this story, a significant part is left out–the cereal boxes weren’t the real turning point. A few months after getting another shot at success, once Airbnb had made it into Y Combinator, the founders noticed something: their listings in New York weren’t converting, even though the demand was there. This meant that the properties and spare rooms hosts were renting out were being viewed by users but were never actually booked. Swiftly, they flew to New York themselves, borrowed a camera, and went door to door photographing their hosts’ apartments by hand. As a result, weekly revenue in that city doubled within just a month. This breakthrough wasn’t driven by code or algorithms, but by the founders getting close to their customers to solve the problem by hand.
The first thing that separates startups that are successful versus those who aren't happens to be the opposite of most people's assumptions. We frequently visualize success as a brilliant, well-designed, and scalable idea. But this story and the data say something blunter: the startups that survive their first year are usually the ones willing to do something completely unscalable, whether by hand, poorly, one customer at a time, all with the intention of figuring out what works before they try to make it big.
Here are the four other things winners across every sector (fintech, enterprise software, consumer apps, and hardware) keep having in common.
They find out if anyone wants it before they build the whole thing. CB Insights studied 431 venture-backed startups that shut down in 2023 and found that 43% failed because of poor market fit, where they built something that the market didn’t actually need. Building something that nobody wants is twice as destructive as the second most common reason a startup dies. While running out of cash is cited in 70% of startup failures, CB Insights clarifies that this is
merely a symptom, not the disease. The money runs out because no one is buying the product, not the other way around.
They grow at the speed their foundation can feasibly support. A Startup Genome study of over 3,200 high-growth tech companies discovered that 74% of them failed because they scaled too early, whether by hiring ahead of revenue, expanding into new markets before the first one worked, or adding features nobody asked for. Not a single company featured in that dataset that scaled prematurely ever reached 100,000 users. The companies that actually paced themselves correctly grew roughly 20 times faster than those that tried to rush into success. Like the idiom “slow and steady wins the race,” ventures that are slower and done right beat the ones that are faster and done wrong, by a factor of 20.
They also admit when the original idea was wrong, fast. Before Brex became a fintech company serving thousands of startups, its founders Henrique Dubugras and Pedro Franceschi were trying to build an augmented reality (AR) company inside Y Combinator. A few weeks in, they realized they had no real idea of what they were doing. Instead of forcing it, they went back to what they actually had an idea of doing: payments, after noticing that neither they nor any of their fellow YC founders could get approved for a basic corporate credit card. Stewart Butterfield captured lightning in a bottle twice using this exact approach. First, his early project Game Neverending became Flickr when the photo-sharing feature proved more compelling than the game itself. Years later, his next game, Glitch, failed too, with him claiming that, according to his own math, 97% of people who signed up were gone within five minutes, and the internal chat tool his team built to stay afloat became Slack. He built two multi-billion-dollar companies by knowing when to let go of his original vision and follow where the actual value was.
They build a team that covers what the founder can’t. CB Insights' studies consistently highlight this fatal flaw: startups frequently stall because they lack either the technical expertise to recruit engineers or the commercial insight to run the business side. One founder wrote after his own company folded that he’d been “blindsided” by how hard hiring became once every engineer wanted to work on something flashier. It’s a quiet failure compared to running out of money, but it shows up constantly in founders’ own accounts of what went wrong.
These five things: 1. doing the unscalable grunt work early 2. testing before building 3. pacing growth 4. admitting fast when you’re wrong, and 5. building the right team doesn’t require funding, connections, or a technical background that most don’t have. We often imagine startup founders as someone born with a rare gift, but that myth is purely the result of survivorship bias. We only hear the stories after success is guaranteed. In reality, before the success story, it just looks like a founder in a San Francisco apartment, hot-gluing cereal boxes shut, trying to make it to next week.
