I spent a long time treating breaks as the part of the day where nothing was happening.
The logic seemed sound. Focus blocks were when work got made, breaks were the gaps between them, so the way to get more done was to shrink the gaps. I ran four hour stretches with a coffee refill in the middle and felt virtuous about it. Then around two in the afternoon my sentences would start going in circles, and I would respond by working later, which made the next morning worse.
What changed my mind was noticing that my second block of the day was consistently good when I had walked beforehand and consistently mediocre when I had eaten lunch at the desk reading something. Same work, same hours, same person. The only variable was the twenty minutes in between.
A break determines the quality of the block that follows it, and whether it works depends on whether it unloads the systems you were just using rather than keeping them busy with different content. That single distinction explains most of what people find confusing about breaks, including why a ten minute phone break can leave you feeling worse than a five minute walk.
The popular model says your brain has a battery and breaks recharge it. That framing is not much use, partly because the science on willpower depletion turned out to be shakier than the popular account suggested, and partly because it gives you no way to tell a good break from a bad one. If it is all one battery, any pause should work, and anyone who has taken a ten minute scrolling break knows that is not what happens.
Try this instead. Focused work loads a particular set of systems: you hold a goal in mind, suppress everything irrelevant to it, process language, make decisions, and track where you are in a structure too big for working memory. A break restores capacity when it stops asking those systems to work, and restores very little when it keeps them loaded and only changes the content.
Line up the two cases.
You stop work, open your phone, and start scrolling. You are now processing language at speed, tracking a rapid sequence of unrelated topics, evaluating each one, deciding whether to tap, and doing a fair amount of social comparison along the way. Every one of those is a system that was already working hard. The subject matter changed, the load did not. Ten minutes later the paragraph you had been holding in your head is gone, and the restart costs more than the break gave you.
Now the walk. No text, no sequence of decisions. Your visual system switches from a fixed plane sixty centimetres away to a field of depth and motion. Your body moves, which changes your physiological state. Attention goes diffuse rather than pointed, and diffuse attention is where an unfinished problem quietly rearranges itself. Five minutes of that can restore more than double the time spent scrolling.
None of which makes phones evil. It makes them a bad choice for the job you are asking a break to do. If you are unsure about a given activity, ask whether you are still reading, deciding, evaluating or holding something in mind. If you are, it is not recovery.
| Break activity | Restores | Keeps loaded | |---|---|---| | Walking, ideally outside | Vision, attention, physiological state | Little | | Looking out of a window | Eyes, diffuse attention | Little | | Stretching, lying on the floor | Body, posture, attention | Little | | Making tea without your phone | Attention, some movement | Little | | Reading an article | Very little | Language, comprehension, decisions | | Email and chat | Nothing | Everything, plus new obligations | | Scrolling a feed | Nothing | Language, visual switching, evaluation |
The bottom two rows are worth sitting with, because they are what most people actually do with a break, and they are the least likely to give any of it back.
There is a second job happening in a break that has nothing to do with rest.
When you switch away from a task, part of your attention stays with it. Sophie Leroy named this attention residue, and it is why a quick check of something costs far more than the seconds it occupies. The mechanism is covered in attention residue and why your brain stays stuck on the last thing, so I will not repeat it here.
The consequence for breaks is this. A break is the window where residue from the previous block either thins out or gets topped up with something new. Spend it on email and you start the next block carrying residue from three messages you did not answer, on top of whatever was left from the work. Spend it walking and the residue fades, which is why solutions turn up away from the desk.
One addition helps. Before you stand up, spend thirty seconds writing down where you were and what the next move is. That closes the loop enough for the background rehearsing to stop, and removes the reorientation cost at the start of the next block.
What kept me skipping breaks was the arithmetic. Four hours minus forty minutes of breaks is three hours twenty, and three hours twenty is less than four hours.
The arithmetic is wrong because the hours are not interchangeable. An hour of fresh attention and an hour of depleted attention produce very different amounts of work, and the fourth consecutive hour without a break sits near the bottom of that range. The real choice is between three hours twenty of decent work and four hours where the last one goes on rereading the same paragraph. Push through often enough and the cost stops staying inside the day. It is a loan against tomorrow's capacity, taken at a poor rate.
The honest answer: it depends on the intensity and length of the block before it, and the research on optimal ratios is thin enough that anyone quoting a precise number is going past what is known.
The widely cited ratios trace back either to the Pomodoro technique, which Francesco Cirillo developed as a personal time management method rather than a research finding, or to one company's analysis of its own time tracking users, which is not a controlled study. Neither is a reason to ignore them. Both are a reason to treat 25/5 and 52/17 as starting points rather than facts about human physiology.
Ranges I would defend:
Intensity matters as much as duration. Ninety minutes of careful debugging or difficult writing costs more than ninety minutes of spreadsheet tidying and deserves a longer break, even though the clock says the same thing.
Take the first two minutes back at the desk as your measurement. If starting feels like turning over a cold engine, the break was too short or the wrong kind. If you drop straight back in, it worked. Crude, but available every single time, which makes it more useful than a precise number you never check.
Then run it properly for a week. Keep the block length fixed and vary only the break: five minutes at the desk, ten minutes walking, fifteen minutes walking, noting how the next block went. Most people find the point where extra break time stops adding anything within four or five days, and it is rarely where the popular ratios said it would be.
The failure that ruins more days than any other is not the skipped break. It is the break that quietly becomes the afternoon.
The mechanism is simple. You stop for ten minutes, open something with no natural end point, and nothing tells you the ten minutes are up. Feeds are built specifically not to end. You look up forty five minutes later and now feel bad as well, which makes restarting harder.
Three things fix it, the first doing most of the work:
Eating at the desk while reading something is the worst of both worlds. You are not resting, because you are reading, and you are not really eating either, in the sense of noticing it. You have also spent the one natural boundary in the working day in the same posture, in front of the same screen, in the same attentional mode as the hours around it.
A proper lunch break is the largest single lever available for the quality of the afternoon. Away from the desk, outside for part of it, no screen. If you work somewhere that reads a visible lunch break as low commitment, take it in two pieces, ten minutes outside and ten minutes eating.
Every point above fails in practice for the same reason. At the moment the break is due you are mid-thought and do not want to stop, so the decision gets made by the part of you least able to judge it.
The fix is to decide in advance and let something else hold the boundary. In Deep Focus I build the day in the Session Planner as an explicit sequence of focus and break blocks, setting each break's length to match the block before it rather than making them all the same. With auto advance on, the transition happens without me negotiating about it, so breaks stop being the thing that gets cut when work feels urgent, which is exactly when cutting them costs most. For a single session the timer does the simpler version: a fixed end and a clear signal that the block is over. Run it on the break too.
One honest limitation. None of this helps if the break itself is the problem. An app that ends your focus block cannot make you walk outside instead of picking up your phone, and the phone is a separate device with its own gravity. That part stays yours.
Twenty to thirty minutes, with movement in it, ideally not in the same room. Shorter breaks work fine after 25 to 50 minute blocks, but a demanding 90 minute stretch leaves more to unload and a five minute pause will not touch it. Expect the block after a long one to be weaker regardless of how good the break was.
It comes from one company's analysis of its own time tracking data, not a controlled study, and it has been repeated far more confidently than the source supports. Seventeen minutes is a reasonable break after about an hour of work, so it is not bad advice. Treat it as a starting point rather than a discovered law.
Most likely because the break kept the same systems working. Scrolling, email and reading all use the attention, language and decision machinery you were trying to rest, so you come back with the same load plus whatever new obligations arrived. It can also mean the break ran long enough for attention to fully disengage, which makes restarting expensive.
Not at the exact moment it is going well. Stopping a good stretch to obey a timer is usually a mistake, and that is a real weakness of rigid interval methods. But flow is not cost-free, the bill just arrives later, so take the break at the next natural seam rather than deciding you no longer need one.
Stand up, move, and look at something far away. That covers the systems most loaded by desk work and it fits in five minutes. Daylight helps. The main rule is negative: no reading, no decisions, nothing without a natural end point.
The block that follows a real break is noticeably better than the one that follows a scroll, and it is not close.
Take one properly tomorrow afternoon, and pay attention to the first two minutes back.
Get the latest productivity tips and Deep Focus updates delivered to your inbox
