For about two years my entire strategy for building concentration was to set a two hour timer and lose.
I would start it and genuinely work for maybe twenty five minutes. Then the restlessness arrived, I would hold out another ten, and around the forty minute mark I would check something, then something else, and the remaining eighty minutes were a slow dissolve. I ended most of those sessions convinced my attention was permanently broken.
It was not. I was walking into a gym, loading a weight I had never lifted, failing, and concluding I had bad genetics.
Focus stamina improves the way most physical capacities improve: sessions slightly harder than comfortable, repeated consistently, with real recovery in between. People get two things wrong. The starting point is almost always set above their current ceiling instead of just below it, and the recovery usually is not recovery. Underneath both sits a subtler point: what you are training is not attention. It is tolerance for the discomfort of staying on one thing.
You have probably seen the claim that human attention spans have dropped below a goldfish's. It does not hold up. The number traces back to a statistics aggregation site rather than a study, the sources it cited do not contain the finding, and journalists who went looking for the original research did not find it. The goldfish half has nothing behind it either.
I am not going to hand you a replacement number, because I do not have an honest one. But here is an observation available to anyone: the same people who supposedly cannot concentrate for eight seconds watch three hour films, play games all evening, and read novels on holiday. A general collapse in the capacity to sustain attention is not a good description of what is happening.
What is real, and more useful, is that attention is heavily context-dependent, that most of us have trained on inputs delivering novelty every few seconds, and that this makes work with slow reward feel much worse by comparison. Different problem, different training.
Watch what happens at the moment a session breaks down. For me it is almost never "I cannot understand this." It is a small itch of understimulation, arriving right when the work turns effortful or ambiguous, and the urge to check something is the escape from it.
This changes the target. If the problem were raw attention capacity, you would train by concentrating harder, which does not mean anything. If the problem is tolerance for a mildly uncomfortable state, you train by staying in it slightly longer than is comfortable, repeatedly, which you can actually do.
The impulse to check your phone is not steady pressure building until you give in. It comes in waves: it rises, peaks fairly quickly, and subsides if nothing feeds it. The reason this is not obvious is that we relieve the urge within seconds of noticing it, every time, so we only ever experience the rising part.
The practical version: when the urge arrives, neither fight it nor obey it. Notice it, keep your hands on the work, let it sit there. Usually it flattens out within a minute or two.
Two things help:
Every urge you sit through is one repetition. That is the entire training stimulus. Sessions where you never felt the pull were pleasant and built nothing.
The standard mistake is picking a session length from an article instead of from your own behaviour. Somebody reads that deep work happens in 90 minute blocks, sets 90, fails, and takes the failure as evidence about themselves.
Progressive overload needs an accurate baseline, and focus has an unusual property here: start clearly below it. In the gym you train near your limit. With focus you are building a behaviour you will repeat daily, and the repetition matters more than any single session. A block you finish is a win logged. A block you abandon teaches you that blocks get abandoned, and that lesson is expensive.
To find your baseline: start a stopwatch, work on something real until the first moment you genuinely want to stop, note the time, stop. Do that three days running and take the lowest of the three, not the average and definitely not the best day. Your first week runs at about 80 percent of it.
If your baseline is 25 minutes, train at 20. Yes, even though you can do 25. Week one is not for pushing, it is for building a run of completed sessions.
Assume a baseline of 25 minutes, roughly where a lot of people land. Scale the numbers to yours and keep the shape.
| Week | Session length | Sessions per day | Break between | Goal | |---|---|---|---|---| | 1 | 20 min | 3 | 7 min | Complete every session. Length is irrelevant this week. | | 2 | 30 min | 3 | 8 min | First real discomfort. Sit through at least one urge per session. | | 3 | 40 min | 2 to 3 | 10 min | Notice which time of day holds up and which does not. | | 4 | 50 min | 2 | 12 min | One clean 50 minute block is the whole target. |
Rules that matter more than the numbers:
Beyond week four, returns thin out somewhere in the 50 to 90 minute range, which conveniently lines up with when work naturally wants a pause. Longer is not automatically better. Two clean 50 minute blocks beat one ragged two hour one.
Spend a 10 minute break scrolling a feed and you have not rested the system you were using. You kept it loaded and changed the content. Language processing, rapid visual switching, social evaluation, decision after decision: the same machinery with better lighting. Then you return and cannot understand why you feel worse than when you stopped.
A break that restores capacity has some of these: no screen, movement, low decision load, ideally daylight. Walking is close to ideal. Staring out of a window is underrated. Tea, stretching, lying on the floor.
My test: if the first two minutes back feel like restarting a cold engine, the break was not a break. After a walk, the restart is nearly free.
Sleep is the largest input and the least popular. A short night costs more sustained attention than anything in this article can add back. If you are averaging five or six hours and trying to build concentration through session design, you are working on the wrong variable, and you will feel it in exactly the part of the session this plan is trying to extend: the second half.
Memory is a bad instrument here. Bad sessions feel longer than they were, good ones vanish, and after two weeks you have a vague sense that things are "about the same," which is the feeling that makes people quit.
The measurement layer does not have to be sophisticated. A timer for fixed blocks, a stopwatch for open-ended sessions where forcing a stop would hurt the work, and a record of what actually happened. In Deep Focus I watch average session length over a date range, because it is the closest single number to what this plan is trying to move, and the focus heatmap, which is better than my intuition about which days and times hold up. Distraction tracking earns its place too: the apps that kept surfacing during blocks tell you what to add to the profile next week, and that list is rarely what you would have guessed.
One number matters most: the proportion of sessions you complete. High, with length flat for two weeks? Add time. Dropping? You are already past your ceiling.
Here is the honest limitation of every plan like this, including this one. Capacity is not stable. It moves with sleep, illness, stress, grief, weather, and whatever happened in your last meeting. A rigid four week ladder assumes a person who does not exist, and the first genuinely bad day presents itself as evidence that the whole thing failed.
It did not. The plan just needs a lower gear, and almost none ship with one.
Mine is a fixed floor: on a bad day I do one 20 minute block and stop. That is the entire obligation. Not a reduced version of the plan, and not "do what you can," which is too vague to bind. One block, at a length I can complete even at 40 percent.
This works because what is being protected is not the hours. It is being someone who does focused sessions. Skip entirely and the next day is harder to start, and the one after that harder again. Do the 20 minutes and the chain holds; pick the progression back up tomorrow.
Expect a bad week in every four to six. That is not pessimism, it is what your own heatmap shows once you keep one.
Most people notice a real difference in three to six weeks of consistent daily practice, with early gains arriving faster than later ones. Going from 20 to 45 minutes is a matter of weeks; 60 to 90 is slower and, for a lot of work, not obviously worth chasing. Consistency drives the curve far more than intensity.
It is a useful analogy, not a literal mechanism. Concentration does improve with graded practice and degrade without recovery, which is why the training framing works. But nobody has established a single depletable "attention muscle," and the better explanations point at habit formation, environmental cues, and tolerance for discomfort rather than a fuel tank you empty and refill.
It is a good starting structure and a poor ceiling. 25 minute intervals help when your baseline is short, because they make sessions finishable, but the fixed 25 becomes a limit once you can hold 40 or 50, and the mandatory break can chop a session exactly when it was going well. Use it as week one scaffolding, then extend the blocks.
Because they supply constant feedback, escalating novelty, and unambiguous goals, so they never generate the understimulation that breaks work sessions. That is not evidence of broken discipline. It is evidence that your capacity is intact and currently only shows up when the task supplies its own reward, which most real work does not.
It helps, because it converts a decision you have to win repeatedly while depleted into one you made once in advance. Same reason people keep the biscuits out of the house rather than relying on resolve at 9pm. It does not train tolerance by itself, so the useful combination is blocking the easy escapes while deliberately sitting through the urges that remain.
The most useful thing I stopped doing was comparing my twenty five minutes to somebody else's four hours.
They started somewhere too, probably lower than they admit, and the ones who got there did it by finishing a lot of small sessions rather than surviving a few enormous ones. Pick a length you can complete today, complete it, and make the next one slightly harder.
Get the latest productivity tips and Deep Focus updates delivered to your inbox
