I spent years waiting for the free weekend that would let me make real progress. The weekends came, I was tired, and I made almost nothing. Then I started doing twenty minutes on weekday mornings and finished more in three months than in the previous three years.

The maths is boring and undefeated: small and often beats big and rare, because small and often actually happens.

The last ten per cent of a project has no reward built into it. The idea is no longer exciting, nobody is watching yet, and everything left is small, fiddly and unglamorous — the naming, the edge cases, the copy nobody will read.

What helps

• Writing down what "done" means before the excitement wears off
• Doing the boring last bits first thing, when the day still has fuel in it
• Accepting that the finished version will be less interesting than the imagined one

That last one took me a decade. Everything you finish is a downgrade from the version in your head, and shipping it anyway is the whole job.

There is a particular kind of productive-feeling day where I change fonts in my editor, try a new note system, and reorganise a folder structure that was fine. At the end of it I have made nothing.

The test I use now: at the end of the day, is there a thing that exists which did not exist this morning? Tool-fiddling never passes it.

The internet rewards the launch, the streak, the overnight thing. It has nothing to say about year three — the quiet stretch where you are just doing the work, no applause, no milestone, the plot line flat.

Year three is where it starts to work. Not because anything changed on the outside, but because you did. Play the long game. Play it quietly. Let the compounding do the loud part.

I used to think the scarce resource was talent. It is not. Talent is everywhere and mostly wasted. The scarce resource is attention — the ability to point yourself at one difficult thing and stay pointed at it long after it stops being fun.

Everyone can do it for an afternoon. Almost no one can do it for a year. That gap is where all the good work lives.

I once shipped something I was not proud of, on a deadline, with bugs I knew about. I braced for the reckoning. It never came — instead, real people used it, told me what was actually wrong (not what I feared was wrong), and I fixed the right things for once.

Shipping badly beats polishing privately, every time. The work does not get good in the drawer.

Working in public is mostly the practice of being slightly embarrassed on purpose. You show the rough thing, the half-formed idea, the draft that is not ready — and the sky does not fall. Someone points out the flaw you were hiding from. The work gets better faster than it ever did in private.

The audience is not the point. The accountability is. Publishing turns "someday" into "this week".

I have tried every productivity system and abandoned all of them except one rule, which is not really a system: protect the first two hours.

Why the morning

Not because mornings are magic. Because they are the only hours the world has not got to yet. No replies owed, no small fires, no one else's priorities in my inbox. Two hours of that is worth six of the afternoon kind.

What they are for

The hard thing. The one that needs a running start and a quiet room. If I spend the two hours on it, the day is already a success and everything after is a bonus. If I spend them on email, no amount of the rest of the day buys them back.

The whole rule

That is it. Protect two hours. Spend them on the thing that matters. Let the day be mediocre after that if it needs to be — it usually does, and it usually does not matter.