Guide
There is no researched optimum. Here is what the evidence does support, and how to pick an interval by testing rather than guessing.
The best interval is the longest one you can finish without cheating, on the kind of task you are doing. For work you have been avoiding, that is often 15 or 20 minutes. For work you find interesting, it may be 50 minutes or more. Nobody has found a number that is right for everyone, and the people claiming otherwise are usually quoting one of two sources that do not support the claim.
Francesco Cirillo was a university student trying to study without drifting. He used a tomato-shaped kitchen timer, tried intervals from about 5 to 30 minutes, and settled on 25 because it worked for him. He has never claimed more than that, and in his own writing the 25 is a starting point you are expected to adjust. There is no experiment behind it, no measured attention cycle, and no reason to think your brain has an opinion about the number.
The number stuck because it is memorable and because a technique needs a default. That is a fine reason for a default and a bad reason to treat deviation as failure.
The other number people quote is 52 minutes of work and 17 of break. It comes from a 2014 blog analysis by DeskTime, a time-tracking company, which looked at the top 10% of its own users as ranked by the app's productivity score and reported the average work and break pattern of that group. That is a correlation inside a self-selected sample of people who chose to install time-tracking software, measured by a proprietary score. DeskTime's own later data revised the ratio upward, which is what you would expect from a statistic that describes a population rather than a mechanism.
None of that makes 52/17 a bad interval. It makes it an interval, on the same footing as 25/5. If a longer block suits your work, the comparison in the Pomodoro versus 52/17 guide goes through the trade-offs.
Two findings are solid enough to plan around. The first is that breaks help. Ariga and Lleras (2011) showed that brief pauses during a long vigilance task prevent the usual decline in performance. Their account is that attention drops because the goal stops registering as new, and a short break refreshes it. What their work does not do is name an interval. The mechanism says take breaks, not take them every 25 minutes.
The second is the one randomized comparison of these methods. Smits, Wenzel and de Bruin (2025, Behavioral Sciences) ran 94 students through two-hour study sessions using either self-regulated breaks, strict Pomodoro, or Flowtime. Strict Pomodoro produced faster-rising fatigue, and both structured conditions showed motivation falling faster than self-regulated breaks. There were no differences in productivity, task completion or flow. One trial with students on a two-hour session is not the last word, but it is the best direct evidence available, and it does not support the idea that a fixed interval extracts more work out of you.
You may also see the interval question justified by ego depletion, the model of willpower as a fuel tank that drains and needs topping up. That model failed two large multi-lab replication attempts, one with 23 labs in 2016 and one with 36 labs in 2021. Fatigue is real; the tank is not a good description of it, and it should not be your reason for picking a number.
The main thing a timer buys you is a cheaper start. If a task makes you reach for your phone, the correct block is short enough that beginning it feels trivial. Twenty minutes for tax paperwork you have postponed twice is not a failure of discipline; it is the length that gets the paperwork started. You can raise it later.
Some work has a long ramp. Debugging a system you have not touched in a month, or writing an argument you have to hold in your head, can take ten minutes just to load the context. A 25-minute block spends 40% of itself on warm-up. For that kind of work, 45 to 90 minutes is more sensible, and the break should be proportionally longer.
If you are in an open office or at home with other people, a long block that keeps getting broken produces a lot of voided sessions and a bad mood. Short blocks survive noisy environments better because they finish before the next interruption arrives.
The common ratios sit between one to five and one to three. Five minutes after 25, ten after 50, seventeen after 52. Longer work blocks need longer breaks, and a break shorter than about five minutes rarely produces the mental gap that makes the next block feel fresh. The ChronoClock timer ships with 25/5, 50/10 and 52/17 as presets and accepts any custom pair, plus a longer break every fourth session.
Run one interval for a full week before judging it. A single day tells you about that day. What you are looking for is not how the block felt but two concrete things: how often you completed the block without breaking the rule, and whether you were still willing to start another one in the afternoon.
If you finish blocks comfortably and want to continue when the timer rings, add ten minutes next week. If you keep cutting blocks short or find yourself dreading them, subtract ten. Two or three weeks of this converges on something that fits your work better than any published number will. Keep in mind that the answer can differ by task type, which is fine. A common pattern is short blocks for administrative work and long ones for the main project.
One more thing worth tracking: total focused time per day. It is easy to tune the interval and miss that you are only running two blocks. Ericsson's research on elite performers found that even they rarely sustain more than about four hours of genuinely effortful practice a day, so four hours is an upper bound, not a target you failed to hit. If you are getting 90 focused minutes on a workday with meetings in it, the interval is not your problem. Use the daily planner to see where the rest of the day is going.