A 90-minute focus session is one long block of uninterrupted work followed by a break of around 20 minutes. It suits work with an expensive warm-up, where you’re still assembling the problem in your head at minute twenty. Three of these is a serious day of deep work.
The ultradian rhythm claim, handled honestly
You’ll see 90 minutes justified by biology. The claim goes: Nathaniel Kleitman, who co-discovered REM sleep, proposed in the 1960s that the roughly 90-minute rhythm of sleep continues through the waking day, cycling between higher and lower arousal. He called it the basic rest-activity cycle. Tony Schwartz built the popular work version of it, including a 2011 Harvard Business Review piece advocating 90-minute blocks.
Some of that holds. Hormone release does pulse on roughly 80 to 120-minute periods during waking hours, and several performance measures fluctuate on similar timescales.
The part that doesn’t hold as cleanly is the jump from “there is a rhythm” to “your brain needs a break every 90 minutes”. A study looking specifically for a 1.5-hour rhythm in cognitive performance found no significant 90-minute periodicity, with longer cycles explaining more of the variance instead. Nothing in the research prescribes a 90-minute work block and a 20-minute service interval for every person.
The other supporting story is K. Anders Ericsson’s work on elite violinists, who practised in sessions of roughly 90 minutes with breaks between them. That’s a real observation about how expert performers organise deliberate practice. It’s also an observation about people who had already found what worked for them, and Ericsson’s broader deliberate-practice framework has been challenged on other grounds.
So use 90 minutes because it fits the shape of your work. Not because of a brain cycle. That’s a weaker argument and a better one, because it’s testable on yourself in a week.
Who these suit
Work with a long warm-up. Unfamiliar code, a difficult proof, restructuring a long document, complex design. If you’re routinely twenty minutes in before you’re any good, a 25-minute session gives you almost nothing.
Writing at length. Not editing a paragraph, actually producing several thousand words of first draft.
Research and reading. Anything where you’re holding several sources in mind at once.
They don’t suit people whose problem is starting. A 90-minute commitment is enormous on a day you’re avoiding the task, and you will simply not press start. If that’s you, 25/5 is the right tool and this page is the wrong one.
They also don’t suit days with meetings scattered through them. A 90-minute block needs 90 clear minutes, and a block interrupted at minute 50 is worse than two 25-minute sessions would have been.
A schedule you can actually run
Three sessions of 90/20 come to 5 hours 10 minutes. That’s a full day’s deep work and it leaves the rest of the day for everything else.
8:30 to 10:00. Session one, hardest work of the day. This is where most of your real output comes from and the only slot you should defend absolutely.
10:00 to 10:20. Break. Outside, if possible. Not email.
10:20 to 11:50. Session two, still on demanding work.
11:50 onwards. Long break and lunch. Take it properly.
13:30 to 15:00. Session three. This one is often better spent on a different kind of task than the morning, because a third stretch of the same problem tends to produce work you’ll redo.
After 15:00. Meetings, email, admin, everything else. It’s the right time for it.
Two is plenty for most people. If you get three good ones, that’s a strong day by any standard, and you should stop rather than look for a fourth.
Making it survive
The 20-minute break matters more here than at shorter intervals, because the fatigue after 90 minutes of concentration is real and doesn’t clear by stretching. Eat, walk, get daylight. Don’t answer anything.
Turn notifications off for the whole block rather than trying to resist them. Ninety minutes is long enough that willpower stops being a plausible strategy.
Split the block if you finish early or run out. There’s no rule that says you must sit there. If session two stalls at minute 60, stop, break, and decide whether to start a fresh one.
Use the 90-minute timer or the full screen timer if you want the countdown out of the browser tab and somewhere you can see it. If the fixed 90 doesn’t fit, flowtime lets the session end when your attention does.
FAQ
Are 90-minute focus sessions backed by science?
Partly. There’s decent evidence for ultradian rhythms of roughly 80 to 120 minutes in hormones and some performance measures, but a study looking for a 1.5-hour rhythm in cognitive performance specifically found none. Use 90 minutes because it matches your work, not because of a prescribed brain cycle.
How long should the break be after 90 minutes of work?
Around 20 minutes, and spent away from screens. After that much sustained concentration a 5-minute stretch doesn’t clear the fatigue, and you’ll feel it in the next block rather than the current one. Food, daylight and movement do more for the next 90 minutes than anything you can do at the desk.
How many 90-minute blocks can I do in a day?
Two is realistic for most people, three is a strong day, four is rare and usually means the later ones weren’t really deep work. Three sessions of 90/20 take 5 hours 10 minutes, which leaves plenty of day for meetings and admin.
Is 90 minutes better than the pomodoro technique?
Better for work with a long warm-up, worse for starting trouble. If you can’t get going at all, a 90-minute commitment is the hardest possible ask and 25 minutes is the easiest. Match the interval to which problem you actually have.
What did Kleitman actually find?
Nathaniel Kleitman identified a roughly 90-minute cycle in sleep and proposed in the 1960s that the same rhythm continues while awake, alternating higher and lower arousal. He did not study office work, and the extension to focus and productivity is an interpretation others added later.
Block out the morning and start: open the 90-minute timer, or run a 90/20 cycle with three sessions.