How long should a focus session be?

Every focus app ships a default, and the defaults disagree. Twenty-five minutes, fifty, ninety, two hours. They are not measuring the same thing, and picking between them is easier once you know what each one is actually claiming.

The 25-minute session is a starting device

The pomodoro number is not a claim about attention. It comes from a technique built around a kitchen timer, and its real purpose is to make starting cheap. Twenty-five minutes is short enough that you cannot reasonably tell yourself you are too tired, which is exactly what it is for.

Use it when the problem is starting. It is the right length for a task you have been avoiding for three days, for the first session of a morning, and for the day you are running on five hours of sleep. It is the wrong length for work that needs a run-up: if the first fifteen minutes are spent reloading a problem into your head, a twenty-five minute session buys you ten minutes of actual work.

The 90-minute session is a claim about depth

Ninety minutes is the number people reach for when the work has a warm-up cost — writing, debugging, design, anything where you hold a structure in your head. The argument is that the expensive part is the reload, so you want fewer, longer runs rather than more, shorter ones.

The trap is scheduling ninety minutes you will not finish. An abandoned session is worse than a shorter one you completed, and not only psychologically: if you are building a habit on a streak or a daily goal, a session you keep failing teaches you that the whole system is optional.

The honest answer: the longest one you will finish

Pick the length you have a better than even chance of completing today, given the day you are actually having. Then lengthen it when finishing gets easy.

That sounds like a compromise. It is closer to a rule about what a session is for. A session is a commitment you make to yourself in advance, and the only property that matters is that the commitment gets kept. A ninety-minute session finished twice a week is a worse habit than a fifty-minute session finished five times, even though the total is nearly the same, because the second one is teaching you something true.

Three practical rules

Match the length to the reload cost, not to the task's size. Answering forty emails is two hours of work with almost no reload cost, so it splits into four sessions with no loss. Untangling one bug is forty minutes of work with a fifteen-minute reload, so it does not split at all.

Let a good session run long rather than planning for it. Extending a session you are inside is a decision made with information — you know you are in flow. Planning two hours on Sunday evening for a Tuesday morning is a decision made with none.

Do not shorten a running session. If the length can be trimmed down to the minutes you have already spent, it was never a commitment; it was a suggestion with a countdown. This is why Deep Focus lets you add time to a running session and never lets you take it away.

Where to start, concretely

If you have no idea: fifty minutes. It is long enough to survive a reload and short enough to be finished on an ordinary day, which makes it the length most likely to still be running in three weeks.

Then let your own record tell you. If you are finishing nearly everything you start, go longer. If you are abandoning more than one session in four, you are not undisciplined — you are committing to a length that does not fit the days you actually have.