The Pomodoro Technique, Honestly Evaluated: What the Studies Show
What two break-timing studies and a micro-break meta-analysis found about Pomodoro, which tasks suit 25-minute intervals, and a two-week test to run yourself.
Ask this guide
Answers come straight from this guide’s text. Nothing is generated, and nothing you type leaves your browser.
Picture two people using the same 25-minute timer. One finds it the only thing that gets her to open the tax folder. The other finds that it rings in the middle of a debugging session and leaves him rebuilding the whole picture from scratch. Both may be right: a fixed interval may help with some jobs and cost you on others, although no study we found tested that directly, and the research behind it is thinner than the technique's reputation.
Short answer: in the two break-timing studies we could verify, fixed breaks produced output similar to self-chosen breaks, with better mood in one and no overall difference in the other. By our reading, not by trial evidence, a timer helps most with starting avoided work and least inside deep, model-heavy tasks.
What Cirillo's method contains beyond a 25-minute timer
Francesco Cirillo began developing the technique in the 1980s, while preparing for a university examination. His own account starts with a small challenge: concentrate for two minutes, timed by a tomato-shaped kitchen timer. Pomodoro is Italian for tomato.
The popular version reduces the method to a routine: work 25 minutes, rest 5, repeat four times, then take 15 to 30 minutes, as a 2025 study summarises it. The official site describes more. It calls the technique "much more than a sequence of 25-minute sessions", treats breaks as "part of the practice, not a reward for doing more", and asks you to record interruptions and your actual effort, then use the records to make estimates more realistic.
The site does not say why 25 minutes was chosen, and we found no research that derives the figure. The studies below test only the timer and break schedule (the meta-analysis, short breaks in general), not the estimating habit, which works like the planning-fallacy correction in our time-blocking guide.
What the studies found about timed breaks
| Study | What was tested | What it found | Main limits |
|---|---|---|---|
| Biwer et al., 2023 | 87 Dutch university students, one day of real study: self-chosen breaks (n = 35), 24 min work/6 min break (n = 25), 12/3 (n = 27) | Self-choosers worked and rested longer, with more fatigue and distraction and less concentration and motivation; task completion and mental effort did not differ, so similar work took less time | One day only |
| Smits, Wenzel and de Bruin, 2025 | 94 analysed students, one 2-hour online session: self-chosen breaks (n = 25), four 25/5 cycles (n = 36), Flowtime (n = 33) | No overall differences in motivation, productivity, fatigue, task completion or flow. After correction, motivation fell faster with Pomodoro than self-chosen breaks, and late-session fatigue was higher with Pomodoro than Flowtime | Unmonitored; single-item ratings; 93% women, all psychology students; no long break after block four |
| Albulescu et al., 2022 | Meta-analysis of 22 samples (N = 2,335), breaks of 8 seconds to 10 minutes | Small gains in vigor (d = 0.36) and fatigue (d = 0.35); performance gain not significant (d = 0.16) | 4 of 22 studies low risk of bias; vigor and fatigue self-reported |
Fixed breaks improved mood in one study and output in none
- The case for fixed breaks is one day-long study plus one null result. In Biwer et al., both fixed schedules produced similar work in less time, with less fatigue and distraction, while self-choosers ran longer sessions (an association from one day, not proof that choice causes tiredness). In Smits's study of 94, nothing differed overall, and its authors conclude that none of the three techniques should be recommended over the others. Neither tracked procrastination or output over weeks.
- Micro-breaks change how you feel more reliably than what you produce. The meta-analysis tested breaks of up to 10 minutes, so it tells you whether to pause, not how long to work. Performance was not significantly better overall. By task type it was significant for creative tasks (d = 0.38) and clerical tasks (d = 0.56, two studies, 95% CI 0.01 to 1.12, flagged by the authors for cautious reading), not for cognitive tasks (d = −0.09).
A 2025 review sounds stronger than its evidence
A 2025 scoping review in BMC Medical Education mapped 32 studies (5,270 participants) and concluded that time-structured Pomodoro intervals improved focus and reduced fatigue compared with self-paced breaks. A scoping review maps what exists; it does not test whether a method works. Its randomised evidence is three trials with 87 participants in total, the same number as Biwer's study, so it appears to be that study counted again. That reading is ours; we did not check its tables.
Variants: Flowtime, long blocks, and the 52/17 figure
| Variant | Rule | What backs it |
|---|---|---|
| Flowtime | You end the block when focus fades. The 2025 study gave 5 min after up to 25 min of work, 8 after 25–50, 10 beyond | Same 2-hour study: no overall difference |
| 50/10, 90-minute blocks | Longer fixed blocks, ended by a timer | No controlled test found; conventions |
| 52/17 | Ratio reported for one vendor's most productive users | DeskTime company data, observational |
DeskTime's 52/17 figure comes from a 2014 analysis of its top 10% of users, ranked by tracked time in apps it classed as productive; they worked about 52 minutes, then took about 17. A 2021 re-run found 112/26, and February 2020 data, which DeskTime notes predates the pandemic, gave 80/17. No sample sizes are reported and the page does not say how breaks were detected. The figure describes highly rated users of one tracker, not what makes anyone more productive.
Why a timer helps you start, and where that claim stops
Piers Steel's 2007 meta-analysis of procrastination research lists task aversiveness, task delay (how far off the payoff sits), self-efficacy and impulsiveness among strong, consistent predictors. A 25-minute box acts on the first: the job becomes "25 minutes of this", not "this, however long it takes", and the bell is a defined way out.
That is our reading; we found no trial testing a timer against procrastination itself, and the studies above did not measure how long people took to begin.
Use the timer as an entry ticket: commit to one interval, and at the bell you may stop without penalty. An if-then plan fixes the cue: "If it is 09:30 and the document is open, then I start a 25-minute timer and write the worst possible version of section 1."
When fixed intervals cost more than they give
Work that depends on a big mental model
Parnin and Rugaber analysed about 10,000 recorded programming sessions from 85 programmers (IDE logs from 2005), splitting sessions at 15 minutes of inactivity. Among sessions with editing, only about 10% had coding start within a minute, and in the 12-developer Visual Studio subset only 7% went straight to editing without navigating elsewhere first. Those gaps were longer than a Pomodoro break and their cause is unknown, so the figures show what returning to a half-built mental model costs, not what a bell costs. Still, a timer is an interruption you scheduled, and whoever returns must rebuild context.
One cue helped. In the roughly 17% of Visual Studio sessions where the first change landed in the method last edited, 35% resumed within a minute: thin support, from 12 developers, for the resume line in the Tip below.
In Smits's study flow did not differ between groups, so the cost did not show up there. We would expect it to be larger for debugging, tracing a data flow or drafting a long argument; that is our expectation, not a finding. The attention residue guide covers the underlying cost.
Days chopped up by meetings
If your open gaps are under 60 minutes, a timer inside them adds ceremony without protecting anything. Treat the gap as the interval and give it one named task; the time-blocking guide shows how to label gaps under 60 minutes. Say a 45-minute gap sits between two meetings: 5 minutes to reload from your last note, 30 on one named task, 10 to write the next note and buffer.
Breaks spent on the inbox
A break filled with work chat, email or news starts another task before you have finished the first. That is our reasoning, not a tested result: Smits's team notes that break activity was neither controlled nor recorded, and research they cite finds that relaxation and socialising can make breaks more effective. Leave the work screen, and talk to a person if you want company.
Tip: Treat the bell as a checkpoint, not a guillotine. If it rings mid-thought, write one resume line (what is done, what is open, your next action) and take the break. If you are inside a hard step, extend once by up to 10 minutes and log the overrun. This is our rule, not a tested one.
Which interval and break to use for which task
The interval column is our starting point; no study we found identifies an optimal length for any of these.
| Task type | Interval | Break | Evidence behind it |
|---|---|---|---|
| Something you are avoiding (tax folder, first draft, dreaded email) | 10–25 min, stopping at the bell allowed | 5 min, off screen | Aversiveness predicts procrastination (Steel); no timer trial |
| Routine queues (inbox, expenses, ticket triage) | 25–30 min | 5 min, stand and move | Breaks up to 10 min helped clerical performance in two studies (d = 0.56, wide interval); length untested |
| Studying or reading new material | 24–25 min, or 12/3 when attention is thin | 5–6 min | Biwer: similar work in less time, less fatigue; Smits: no overall difference |
| Idea generation, outlining | 20–30 min | 5–10 min, change scenery | Creative tasks gained (d = 0.38) from breaks up to 10 min |
| Deep, model-heavy work (debugging, design, hard proofs) | 50–90 min or Flowtime; bell is a checkpoint | 10–15 min, away from the screen | Cognitive tasks: no significant gain (d = −0.09); interruption costs in coding; length untested |
| Fragmented day (gaps under 60 min) | The gap itself | Whatever the day allows | Our judgement; see the time-blocking guide |
| Low-energy stretch | 15–20 min | 5 min | Micro-breaks lowered fatigue (d = 0.35, self-reported) |
A two-week experiment to find your own interval
Pick one task type you will do at least ten times in two weeks, such as first drafts, code review or study sessions.
Define two conditions. A is how you work now; B is the interval and break from the table. Keep session length equal, or compare output per hour.
Assign the ten working days five and five, never more than two same-condition days in a row, for example A B B A B A A B A B. Alternating beats "week 1 against week 2", where a bad week gets blamed on the method.
Log every session. "First real work" means the first sentence typed or first line of code written, not opening the file. On both 1–5 scales, 5 is the most: best focus, most tired.
date | cond | first work | desk | output | focus | tired 12 Nov | A | 14 | 90 | 610 | 3 | 4 13 Nov | B | 6 | 90 | 590 | 4 | 3On day ten, compare medians, not means, so one outage day cannot swing the result.
Say you test first drafts, counting words per 90-minute session. These are hypothetical medians over five days each:
| A: no timer | B: 25/5 | |
|---|---|---|
| Minutes to first real work | 14 | 6 |
| Words per session | 620 | 590 |
| Focus (1–5) | 3 | 4 |
| Tiredness (1–5) | 4 | 3 |
Our decision rule, checked in order:
- Keep A if B's output is 10% or more below A's.
- Keep A if B's focus is 1 point or more lower, or its tiredness 1 point or more higher.
- Otherwise adopt B if time to first real work fell 5 minutes or more, or output rose 10% or more.
- Otherwise it is a tie; pick the one you prefer.
Here B has about 5% fewer words with better focus and tiredness, so rules 1 and 2 pass, and B starts eight minutes sooner, so rule 3 adopts it. Every threshold is ours. Ten days of one person's data is noisy, and a routine can feel better simply because it is new, so repeat the comparison after a month.
Where to start this week
- Today: run one 25-minute interval on the task you have avoided longest, free to stop at the bell.
- Mid-week: note the minutes between sitting down and starting real work on three tasks.
- Next two weeks: run the A/B comparison and keep the log.
Frequently asked questions
What should I do in the five-minute break?
No study we cited controlled break content; our rule: anything that does not hand your mind a new task. Stand, drink water, look out of a window. Work chat, email and news start another task, and the leftover thoughts follow you back, as the attention residue guide explains. In the meta-analysis, longer breaks (all up to 10 minutes) went with larger performance gains across studies, a comparison rather than a test, so treat a longer break after demanding work as something to try.
Is 25 minutes the right length?
Nothing we found identifies it as optimal. The studies used 24 and 25 minutes, and DeskTime's reported numbers moved from 52 to 112 over seven years. The official site asks you to adjust your process in small steps, though it does not say whether the 25 itself may change. Start from the table and let the experiment decide.
Do I have to count pomodoros?
Recording effort is part of the method, and a tally of intervals per task type is the simplest record. Write the number you expected beside the number you used, and the gap shows how far off your estimates run. Without records the timer is a bell.
Does the Pomodoro technique work for ADHD?
We found no controlled trial of the technique in people with ADHD, and the two break-timing studies sampled university students. Treat the table as a starting point, not a prescription; the two-week experiment shows which interval and break suit you.
Sources and further reading
- What is the Pomodoro Technique? — Francesco Cirillo, pomodorotechnique.com, 2025–2026 (undated page). Origin, records, interruptions, breaks.
- Understanding effort regulation: Comparing "Pomodoro" breaks and self-regulated breaks — Biwer, Wiradhany, oude Egbrink and de Bruin, British Journal of Educational Psychology, 2023. Sample and results.
- Investigating the Effectiveness of Self-Regulated, Pomodoro, and Flowtime Break-Taking Techniques Among Students — Smits, Wenzel and de Bruin, Behavioral Sciences, 2025. Three-condition comparison and trend analyses.
- "Give me a break!" A systematic review and meta-analysis on the efficacy of micro-breaks for increasing well-being and performance — Albulescu et al., PLOS ONE, 2022. Effects by outcome and task type.
- Assessing the efficacy of the Pomodoro technique in enhancing anatomy lesson retention during study sessions: a scoping review — Ogut, BMC Medical Education, 2025. The scoping review discussed above.
- Does the 52-17 rule really hold up? — DeskTime company blog (page dated 2024; analysis 2014, re-run 2021). Method and caveats.
- The nature of procrastination: A meta-analytic and theoretical review of quintessential self-regulatory failure — Piers Steel, Psychological Bulletin, 2007. Predictors of procrastination.
- Resumption Strategies for Interrupted Programming Tasks — Parnin and Rugaber, 2009 conference version (IEEE International Conference on Program Comprehension), extended in Software Quality Journal, vol. 19, online 2010. Time to first edit after programming breaks.