Being right once only counts if the misses were small
The arithmetic of startups rewards a single large win. It only works for founders who are still standing when that win comes.
Don’t worry about failure; you only have to be right once.
The arithmetic behind this line is the arithmetic of venture capital. A fund that backs twenty young companies expects most of them to fail or return little. It makes its money on one or two that grow very large, and those one or two can repay everything else several times over. Losses are capped at the amount invested. Gains have no ceiling. In that world a string of failures followed by one success is a good record, not a bad one.
Drew Houston knows the pattern from the founder’s side. Before Dropbox he had worked on other ventures, including an online test-preparation business, and he has described the Dropbox idea as coming from a moment of irritation on a bus from Boston, when he realized he had left his USB drive at home and could not get to his files. He started writing code on the trip. The company went through Y Combinator in 2007, and a short demonstration video posted the following year brought tens of thousands of people onto its waiting list almost overnight. By 2010 it had millions of users. The earlier attempts did not matter much once the one that worked arrived.
The line is widely credited to him in this form; no transcript or date for it has been traced, so it reads best as his summary of that experience.
The advice is sound on one point. Fear of failure keeps many capable people from starting at all, or makes them start timidly, choosing ideas so safe that success would be small too. If the payoff structure really is lopsided, that fear is miscalibrated. A founder who has tried three things and failed three times has lost less than it feels, and has learned things that make the fourth attempt more likely to work.
The trouble is the condition the line leaves out. Being right once only helps a founder who is still in the game when it happens. That requires every failure before it to be survivable, and many are not. A founder who funds an idea with personal loans or a mortgage on the family home carries a loss that outlasts the company. A co-founder relationship ruined in the first attempt may not be available for the second. Years spent are gone for good, and a thirty-year-old has more of them to spend than a fifty-year-old.
There are also fields where the logic simply does not apply. A surgeon or a bridge builder is not allowed to be wrong many times on the way to being right once. Some startups share that property. A company that holds customers’ money or runs medical records cannot treat failure lightly, because the cost falls on other people.
So the useful form of Houston’s line has two parts. Worry less about whether an attempt will fail, because one success can outweigh many misses. Worry a great deal about how much each failure will cost, and keep that cost small enough to walk away from. The founders who end up being right once tend to be the ones who arranged their earlier mistakes to be cheap.
A plain way to do that is to decide in advance how much time and money an attempt gets before it is judged. Once the limit is written down, the failure has a known size. That turns a vague fear into a budget, and a budget is something a founder can plan around.