San Francisco, late October 2007. On the kitchen table, a bill. The rent on the apartment is $1,150. Brian Chesky has $1,000 in his bank account (Penguin).
Joe Gebbia, his roommate, sets his laptop down next to the bill.
“Did you see? The world industrial design congress is coming to town. A few thousand designers. The hotels will be full, and overpriced.”
“So?”
“So we have room in the living room. And three air mattresses in the closet.”
Brian looks at him, half convinced: “You want to rent our living room to strangers?”
“With breakfast. We call it Air Bed and Breakfast.”
I am reconstructing the dialogue, but the facts are documented. The two designers throw together a website in a few days. Three people book, at $80 a night. One of them, Amol Surve, stays five nights and pays $400 (Penguin).
Nothing to pop champagne over. Enough to cover part of the rent. And above all enough to get a piece of information nobody had: strangers will sleep in strangers’ homes, and pay for it.
Three customers beat a beautiful storefront
People often ask me: how do you win customers online when you start from zero?
My answer fits in one sentence. You need a real problem, solved at the right moment, for a few people who truly need it solved. The perfect website will come later, if the problem exists.
Look at how precise the problem spotted by Chesky and Gebbia was. One city, one week, and thousands of visitors looking for a bed just as hotel supply maxes out. The shortage had a date and a location. Their solution was imperfect, but it was available immediately.
Here is the irony of the story. These two men were designers. They could have spent three months polishing a visual identity and a slick interface. They did the opposite: a bare-bones site, three mattresses, a breakfast.
Most people who want to sell online follow the reverse order. First the domain name, then the brand guidelines, then the website, then social media. The first conversation with a potential customer comes last, if it comes at all. Six months of work then rest on an assumption nobody has checked.
What the legend leaves out
The version circulating on social media is beautiful. Too beautiful. It contains two errors I need to correct, because they would mislead you about the method.
First error: the idea that they did no promotion at all. In reality, Chesky and Gebbia emailed design blogs and the conference organizers and asked them to help promote their website, which they did (Penguin). No advertising budget, true. But targeted promotion, done by hand, aimed at exactly the people who had the problem. Remember this: your first customers do not fall from the sky. You go and get them where they already are.
Second error: the idea that one weekend of air mattresses was enough to create an income machine. The following months were hard. In 2008, the founders were piling up credit card debt. To pay it off, they designed limited-edition cereal boxes, “Obama O’s” and “Cap’n McCain’s”, in the middle of the US presidential campaign. They sold 500 of them at $40 each, for $20,000 (Inc.).
Joe Gebbia described that period with a piece of startup jargon: the “trough of sorrow”, the moment when a market and a product both exist but fail to fit together (Inc.). Paul Graham, cofounder of the Y Combinator accelerator, had doubts about the idea itself. He accepted the team after hearing the cereal story, convinced by its grit more than by its concept (Inc.).
I know that trough. I founded my first company while I was a PhD student at the IBM Research Centre in La Gaude. Nothing went as planned the first time. Innovation is like sex: nothing works perfectly the first time.
Beware of survivorship bias
Now we need to look at this story with a critical eye. We tell the Airbnb story because Airbnb survived. How many pairs inflated mattresses, rented out a room, pocketed a few hundred dollars, then never found a second customer? Nobody counts them. Nobody writes their story.
This is survivorship bias: we derive a recipe from the winners and forget the losers who followed the same recipe. Three first customers do not predict success. They provide a signal, nothing more.
The numbers are a reminder of how hard the ground is. CB Insights analyzed 431 venture-backed startups that shut down since 2023. For 70% of them, the reason given was running out of capital, and for 43%, poor product-market fit (CB Insights). The firm also points out that running out of cash is almost always the final cause of death, rarely the root cause (CB Insights).
In other words, many companies die from having built, for a long time and at great expense, a product the market did not want. Three air mattresses cost less than a funding round to discover the same truth.
That is why I keep repeating that assumptions are highly dangerous and should be killed with experimentation as early as possible. In discovery methods such as lean startup, the quality you aim for at the start is deliberately imperfect: you are looking for proof before you aim for perfection (My book, chapter 12).
Reverse the order: the five-step method
Here is how I would transpose the mattress story to your situation. Five steps, in this exact order.
- Find a precise problem, experienced by someone, right now. Precise means dated, located, and expressible in one sentence. “People want to travel differently” is an intuition. “Thousands of designers won’t find a room during the last week of October” is a problem.
- Offer a makeshift solution. A simple page, a form, a direct message, a phone call. If you are embarrassed by your first version, you are probably moving at the right pace.
- Serve a handful of people by hand. Talk to them. Host them. Chesky and Gebbia made breakfast for their first guests. You will learn more from one conversation than from ten dashboards.
- Watch what they pay for and what they criticize. Payment validates value. Criticism shows you where to improve. A compliment without payment validates nothing.
- Build what comes next from these real answers, not from your assumptions. Every new feature should respond to an observed behavior.
Rate yourself honestly on a scale of 0 to 5: how many potential customers did you interview before launching your last project? If the answer is zero, you know where to start again.
Do things that don’t scale
In July 2013, Paul Graham published an essay that became a reference for founders: “Do Things That Don’t Scale”. His first observation is crystal clear: the most common unscalable thing founders have to do at the start is to recruit users manually (Paul Graham).
He cites Airbnb specifically. During their time at Y Combinator, the founders went door to door in New York, recruiting new hosts and helping existing ones improve their listings (Paul Graham).
Brian Chesky has recounted the advice Paul Graham gave him at the time: focus on 100 people who love you rather than a million people who kind of like you (Fortune). Those hundred people become your ambassadors. They talk about you without you paying for it.
Transpose this to your business. Your first hundred customers will not come from a campaign. They will come from your personal messages, your calls, your presence in the communities where your problem is already being discussed. It is slow. It is uncomfortable. And it is precisely what your competitors avoid doing.
Build something that belongs to you
One last angle remains, often forgotten in this story. Chesky and Gebbia did not send out résumés to solve their rent problem. They looked at what they had at hand: a living room, mattresses, a network in the design world. Then they defined a goal based on those means.
This way of reasoning has a name: effectuation, theorized by Saras Sarasvathy, a professor at the University of Virginia’s Darden School. Causal reasoning starts from a goal and looks for the means to reach it. Effectual reasoning starts from the available means to define the best possible goal. Two of its principles fit the mattress story perfectly: “the bird in hand”, which favors existing means, and “affordable loss”, which means deciding what you can afford to lose rather than speculating on future gains (My book).
What did the two designers risk? A few days of work and their peace and quiet for a week. An affordable loss. Audacity means doing something whose outcome is far from certain, and you limit its risk by quantifying the loss before acting (My book, chapter 6).
I do want to stay honest with you, though. Building your own system does not mean it will run on its own the next day. Airbnb’s founders went through months of debt and doubt before the company took off. Be wary of passive-income sellers who promise an autonomous machine in a few weeks. What they sell is an assumption, rarely a verifiable result.
The decisive shift happens elsewhere. It comes the day you stop waiting for someone to give you permission to solve a problem.
The essentials
Remember four things:
- Your first customers come from a precise problem, with a date and a location, solved at the right moment, long before the perfect website.
- Even without an advertising budget, you have to go and get your first customers by hand, where they already are.
- Three customers provide a signal, not a guarantee: survivorship bias hides everyone who followed the same recipe without success.
- Start from your means and from the loss you can accept, then kill your assumptions with experimentation.
Customer zero is never born from a perfect marketing strategy. It is born from your ability to see a problem others ignore, and to decide to solve it yourself. Innovate or agonize, the choice is up to you.
References
- (My book) https://philippeboulanger.com/book/
- (Penguin) https://www.penguin.co.uk/discover/articles/the-air-bnb-story
- (Inc.) https://www.inc.com/tess-townsend/airbnb-gebbia-trough-of-sorrow-npr.html
- (CB Insights) https://www.cbinsights.com/research/report/startup-failure-reasons-top/
- (Paul Graham) https://paulgraham.com/ds.html
- (Fortune) https://fortune.com/2023/05/25/airbnb-ceo-cofounder-brian-chesky-best-advice-paul-graham/
