How long should a pomodoro be?

Twenty-five minutes is the default, not the answer. What the research actually says about focus blocks, when 50 or 90 minutes is better, and why the break is the part people drop.

The short version

25/5 is where to start. It is short enough that starting is easy, which

is the problem the method solves.

50/10 once you are warmed up on work with a long ramp, writing, code,

anything where the first ten minutes are spent reloading context.

90/20 for deep work, matched to the body's own ultradian cycle.
15/3 on a bad day, or for a task you have been avoiding. A short block

you actually start beats a long one you do not.

The break is not optional. It is the part that makes the next block

possible, and it is the first thing everybody drops.

Where twenty-five minutes came from

Francesco Cirillo was a university student in the late 1980s, struggling to study, and he made a deal with himself using a tomato-shaped kitchen timer: ten minutes without getting up. Twenty-five is where he eventually settled, long enough to get somewhere, short enough that committing to it is not a decision you have to think about.

That is the whole derivation. It is not a finding about attention span, and nobody should treat it as one. What it is, is a well-chosen default: small enough to beat procrastination, large enough to be worth the setup.

What the research actually supports

Three things, none of which name a number:

Sustained attention decays, and recovers with a break. This is well-established and is the mechanism the method relies on. The decay is gradual and the recovery is fast, which is why short frequent breaks work better than one long one.

Switching costs are real and large. Interrupting a task and returning to it costs minutes, not seconds, and the cost is worse the more complex the task. This is the strongest argument for the method: a block is a commitment to not switch, and the timer is what makes that commitment external.

The break has to be a real break. Checking messages is not a break from concentrating; it is a different task. Standing up, looking out of a window, making tea, those restore. This is where most people get it wrong.

The famous "52 minutes on, 17 off" figure is from a productivity app's analysis of its own users' data, not a controlled study. It is interesting rather than authoritative, and it is offered here as a preset for that reason.

Choosing your length honestly

The question is not "what is optimal" but "what will I actually start". Those are different questions and the second one is the one that matters.

Pick shorter if you are avoiding the task, the work is unpleasant, you are tired, or you have been failing to start for days. Fifteen minutes of something is infinitely more than none of it.

Pick longer if the work has a long warm-up, you are being interrupted by your own timer, or twenty-five minutes keeps ending just as you get somewhere. That last one is the clearest signal: a timer that interrupts flow is worse than no timer.

The test: if you routinely want to keep going when it rings, the block is too short. If you routinely check the clock, it is too long.

The long break, and the four-block cycle

Every fourth block earns a longer break, fifteen to thirty minutes. The number four is another convention rather than a finding, and the timer here lets you set it to whatever suits, because two long blocks of ninety minutes plainly do not want the same rhythm as four short ones.

What matters is that the long break is long enough to actually leave: eat something, walk somewhere, have a conversation. A long break spent at the desk is a short break with extra guilt.

Why the timer has to be external

The reason a kitchen timer worked for Cirillo, and the reason a phone stopwatch usually does not, is that the commitment has to be to something outside your own judgement. "I will work until I feel like stopping" is not a commitment; it is a description.

This is also why a timer that quietly drifts or stops when the tab is backgrounded undermines the whole thing. If you cannot trust the timer, you start checking it, and checking it is the interruption you were trying to avoid. The timer on this site works from the clock time the block ends at rather than counting seconds, which is what makes it correct after twenty minutes in a background tab.

What it will not fix

The method is a way of protecting attention you already have. It does not create it. If you cannot focus because you are exhausted, anxious about something else, or working on something genuinely unclear, a timer will make that visible rather than better, and noticing that is useful in itself.

It also suits some work badly. Anything with a long unbreakable unit (a meeting, a lab procedure, a conversation) does not fit into blocks, and forcing it in is how people decide the method does not work for them.