OurWord.See how people before you handled it.
Contents All

Starting and building

Paul Graham

Do things that do not scale

Paul Graham (United States · 1964–) — Do things that do not scale.

Starting and buildingModernDo things that do not scale

What actually happened?

Y Combinator began as a summer experiment with a deliberately strange shape: fund a whole batch at once, give each of them a small amount of money and three months, and have dinner with them every week. Every part of it was the opposite of how investment worked — small cheques, no track record required, founders as young as they came, decisions in a ten-minute interview. The batch was the point. Founders learn more from the company two tables over, at the same stage and similarly frightened, than from anybody with an office.

Make something people want

It is the only thing on the wall at YC. His judgement after thousands of companies is that the fatal mistake is singular: building something nobody actually wanted. Everything else — the wrong hire, the wrong price, the bad quarter — is repairable.

A team can survive being outbuilt. It cannot survive having been right about a need that was not there.

Schlep blindness

The best openings often sit behind work that is tedious and unglamorous — bank integrations, compliance, refunds. The mind skips over those ideas so quickly that you never notice having skipped. Which is exactly why they are still available.

Everybody saw the market. The reason nobody took it is that the first nine months are paperwork.

A maker's day is not divisible

Managers work in hours, makers need whole blocks. A single meeting in the middle of an afternoon does not cost an hour; it costs the afternoon, because neither remaining piece is long enough to get into the work.

Push every meeting to one end of the day. The output change is larger than any productivity system.

Better a few who love it

A hundred users who love the thing are worth more than ten thousand who find it acceptable. The first group tells you why, brings other people, and forgives the rough edges. The second leaves quietly and teaches you nothing.

Chasing a satisfaction score upward across everyone is usually how a product becomes something nobody would miss.

How do I use it today?

In the cold-start phase do not build channels, buy traffic or automate. Find the first hundred users by hand and serve each one until they are happy; when that work is crushing you, that is the signal to scale. Test an idea on three questions: do I want it, can I build it, and why has nobody done it — no answer to the third usually means it is not worth doing, or you have not yet seen the tedium hiding in it.

Deep read

Do things that do not scaleThe early work is meant to be handmadeThey flew to New York and photographed the apartments themselvesDefault alive or default deadRun this number before anything elseMost founders do not know which one they are

Read alongside

Peter ThielStart from the users in front of you, or start from whether the market can ever hold a profit. Peter DruckerGraham writes about the first hundred users and doing things that do not scale; Drucker writes about what happens after scale, when the activity outlives the reason for it. Kevin KellyDo things that do not scale and a thousand true fans are the two ends of one idea: the first says start by hand, the second says by hand can carry you all the way.

Further

Peter Thiel Naval Ravikant Jeff Bezos Steve Jobs Peter Drucker Kevin Kelly

Lines to keep

There is one fatal mistake: building something nobody wants.

Do things that do not scale. It is the only route to scale.

Better to have a few people love you than many find you acceptable.