The thousand days before Airbnb looked sudden

The count checks out, roughly. In October 2007 Brian Chesky and Joe Gebbia, two designers short of rent, put air mattresses on the floor of their San Francisco apartment and offered them to visitors to a design conference that had filled the city’s hotels. This month the company that grew out of that weekend, Airbnb, announced a funding round of more than seven million dollars led by Greylock. Between the two dates lie a little over a thousand days, and most of them were unglamorous.

The site was launched more than once before anyone noticed it. The founders tied one relaunch to the 2008 Democratic convention in Denver, when hotel rooms were again scarce. When money ran out that fall they designed and sold collectible cereal boxes themed on the two presidential candidates, which paid off their credit card debts and later caught the eye of Paul Graham at Y Combinator, who admitted them to the winter 2009 batch. Even then bookings were thin. In New York, the founders found that listings with poor photographs did not rent, so they flew out, knocked on hosts’ doors and took the pictures themselves.

The value of Chesky’s phrase is in the quotation marks around “overnight”. Outsiders see a company at the moment it becomes visible, usually when it raises money or appears in the press, and they mistake that moment for its beginning. Founders reading those stories compare their own third year of struggle with someone else’s announcement and conclude they are failing. Counting the days is a corrective.

It can also mislead. A thousand days is not a recipe. Plenty of companies spend a thousand days struggling and then close, and nobody quotes their founders. Endurance kept Airbnb alive, but endurance alone did not make it work. Specific decisions did: meeting hosts in person, fixing the photographs, choosing a market where the problem was sharp. The time mattered because it was spent learning, and the learning is what a younger founder should copy.

The days were not evenly hard, either. Most of the progress arrived in the last stretch, once the founders understood what their users needed, and the first two years mainly bought them the chance to find that out. A founder counting days should also ask what each month taught, and whether the next month is likely to teach something new or simply repeat the last one.