Make failure cheap, make learning fast
Originally written in Chinese. Translated with AI, reviewed by Gary. Read the Chinese original →
I recently came across a point Paul Graham makes about VC, and I found it quite interesting.
In ordinary business, our instinct says: the higher the success rate the better, the lower the failure rate the better.
But in venture capital, it isn't that simple.
The big money in VC usually doesn't come from eight out of ten companies doing "pretty well". It comes from ten or twenty companies where most fail, and one or two end up returning tens or hundreds of times the money, or even more.
So if a VC tries hard to push his failure rate down, he will most likely become more and more conservative.
He starts picking deals that look safer, easier to understand, less likely to go wrong.
The problem is that something everyone understands, and everyone expects to succeed, often won't produce a spectacular return.
Because everyone can see it, competition naturally moves in.
The opportunities that produce huge returns often look strange, uncertain, even like a bad idea at the very beginning.
This logic applies to building a business too.
Of course, founders aren't VCs.
A VC can invest in twenty companies and lose fifteen. A founder can't fail his own company fifteen times.
We're not trying to "chase failure".
What we should really chase is having more failures we can afford.
Those two things are very different.
If a new idea needs twenty hires, RM2 million and eighteen months before you know whether it's right, your company might only get one or two tries.
But if the same hypothesis can be tested first with two people, two weeks and a little money, you might be able to try ten or twenty times a year.
Out of ten tries, seven get no response, two are average, and one suddenly produces a completely different result.
That one might be what people call a non-linear opportunity.
Why are so many big opportunities found by trying rather than by thinking?
Because at the start, you have no idea which path will suddenly scale.
All you can do is keep throwing small stones and listening for where the echo sounds different.
The point isn't to hit with every stone.
It's not to use up all your stones on the very first throw.
In a way, building a business is a game of probabilities.
Survive long enough, try enough times, and never let any single mistake knock you out of the game.
Many people say founders need to raise their success rate.
Maybe what matters more is raising the number of attempts.
Not random attempts, of course.
Every attempt should have a hypothesis, a cost ceiling and a clear way to judge it.
If it's wrong, stop.
If there's a bit of response, keep going.
If a strong signal suddenly appears, concentrate your resources there.
The real danger isn't failure.
It's being so afraid of failure that you only dare to do things that "shouldn't go wrong".
Because if an opportunity is so safe it almost can't fail, it usually also means it can hardly give you a very different result.
Building a business isn't about avoiding every failure. It's about making failure cheap, making learning fast, and keeping yourself at the table.
Then keep playing.
Until one day you find an opportunity worth slowly raising your stake on.