Focus sessions that survive a real day

Every focus method I have tried was designed for a calm day. Twenty-five minutes on, five off, repeat. It works beautifully on a Sunday morning and falls apart on a Tuesday, which is the only day that actually matters.

A real day has a courier at the door, a message that cannot wait, a build that fails, and a stretch after lunch where your brain simply refuses. If your system assumes none of that happens, the system is not the thing being tested — you are. And you will lose, because you are up against a day, not a lack of discipline.

I have been rebuilding my own approach around a different assumption: the interruption is not the exception, it is the shape of the day. Three rules came out of that, and they are the only ones that have survived six months.

Rule 1: Plan the session, not the day

Daily plans are a forecast, and forecasts about your own attention are worthless past about 11am. By mid-afternoon the plan is a list of things you did not do, which is a demoralising object to carry around.

A session plan is smaller and honest. Before I start, I write one line:

90 min · Rewrite the onboarding copy
Done when: all six screens have final text in the file

Two things matter here. The first is the duration, which is a budget rather than a promise. The second is the done when, which has to be something you can point at. "Work on onboarding" has no end. "Six screens have final text" ends.

Anything I think of during the session that is not that goes on a scrap list, unsorted, to deal with later. Not a proper task manager — a scrap list. The whole value is that capturing something takes three seconds and no decisions.

Rule 2: Make the first move stupidly small

The expensive part of a focus session is not the ninety minutes. It is the ninety seconds before it, where you are deciding whether to start. That gap is where the whole day leaks away.

So the first action has to be too small to argue with. Not "write the onboarding copy" — open the file and read the last paragraph I wrote. Not "fix the crash" — run the app and reproduce it once. The point is not the task, it is that you cannot negotiate with something that takes eleven seconds.

What surprised me is how often the small move is the whole unlock. You open the file, read the last paragraph, notice the obvious next sentence, and you are twenty minutes in before you remember you were reluctant.

Rule 3: Resume, don't restart

This is the one that actually changed things, and it is the rule most timers get wrong. If a pomodoro is interrupted, the orthodox advice is to void it and start a fresh one. That is a rule designed to protect the purity of the method, not your afternoon.

The honest version is that interruptions are going to happen four or five times, and if each one costs you a whole session you will end the day with zero completed sessions and the belief that you cannot focus. You focused fine. The scoring was rigged.

So I let a session pause and resume. The timer holds. The done when line stays on screen. Coming back to a paused session with a visible finish line is a completely different experience from coming back to nothing and having to decide, again, what you were doing.

What happens	Old rule	What I do now
Two-minute interruption	Void the session	Pause, deal with it, resume
Twenty-minute interruption	Void the session	Pause, resume, re-read the last thing I wrote
Session is genuinely dead	Void the session	Close it out, write one line about where I stopped

That last row matters as much as the others. A session that ends early is still worth closing properly, because the one line about where you stopped is what makes the next session cheap to start.

The break is part of the session

A break spent scrolling is not a break, it is a different kind of work with worse pay. You come back with less attention than you left with, which is a strange thing to accept from something you did to recover.

The breaks that actually restore something have a hard ending built in — a walk to the end of the street, a coffee that finishes, one short game. I have written separately about why a short arcade run works so well here, but the short version is that it ends on its own, and nothing about it follows you back to your desk.

The test is simple: at the end of the break, is there an open thread? A half-read argument, a video queued up, a message you have not replied to? If yes, the break is going to keep charging you rent for the next hour.

What I actually track

Almost nothing, and that is deliberate. Elaborate tracking is a productivity system's way of becoming the work.

• Sessions completed this week. One number. Not hours, because hours reward sitting there.
• What the session was for. The done when line, kept.
• Nothing else. No focus score, no streak, no weekly report I will not read.

Six sessions in a week is a good week for me. Ten is a great one. Seeing that number is enough feedback to keep the habit alive without turning it into a game I would eventually start playing instead of working.

The whole thing, short

1. Write one line before you start: how long, and what done looks like.
2. Make the first action small enough that refusing it would be silly.
3. Pause and resume through interruptions instead of voiding the session.
4. Take breaks that end by themselves.
5. Leave the next first move behind you.

None of that requires an app, though building one is how I ended up thinking about it this hard. If you want the version with the timer and the shared rooms already assembled, that is Lume Flow. If you want the version that costs nothing, it is the five lines above and a scrap of paper. Both work. The paper has fewer bugs.