How to Actually Focus

You have seen the statistic. Every interruption costs you 23 minutes and 15 seconds of recovery. It appears in productivity books, on slides, in the marketing for apps that block other apps.

It is worth going and reading the study, because the most-cited experiment on interrupted work found something different, and the real finding is more useful than the one people quote.

The short answer

Four changes, in order of how much they do.

None of these are clever. The reason they work is that they act on the mechanism rather than on your intentions.

What the study actually found

The paper everyone half-remembers is Mark, Gudith and Klocke, presented at CHI in 2008. Participants worked on an email task and were interrupted by a supervisor with questions, either about the same subject or about something unrelated, and compared against a condition with no interruptions at all.

So the popular version has the sign backwards. Interrupted people in that experiment were not slower. They were faster, and worse off.

That reframes the problem in a way that matches most people’s actual experience. You get to Thursday having produced roughly what you meant to produce, and you feel wrecked, and you cannot point to the reason because the output was fine. The cost was charged to a different account.

Why this matters for how you organise a day

If interruption cost time, the fix would be scheduling. Since it costs effort and stress at a constant quality of output, the fix is different: protect the conditions rather than the calendar.

Three consequences worth taking seriously.

Feeling productive is not evidence. The compensation mechanism means an interrupted day can produce normal output. You cannot detect the problem by looking at what you finished.

Tiredness is data. If you are exhausted after a day of ordinary work, the interruption load is a candidate explanation, and a more likely one than a personal failing.

Removing interruptions may not make you faster. It may make the same work cost less. That is a smaller promise than productivity advice usually makes and it is the one supported here.

The people who do it most are not better at it

The other assumption worth dismantling is that heavy multitasking builds a skill. It was tested directly, and the result went the other way.

The direction of causation is genuinely unknown here, and the authors do not claim otherwise: it is possible that multitasking degrades filtering, and equally possible that people who filter poorly are drawn to multitasking. The samples were also small, in the range of fifteen to twenty-two people per group.

But either reading kills the same idea. Nobody in this study was getting better at handling distraction through practice. Whichever way the arrow points, “I work well with things going on around me” has no support in the one experiment that looked for it.

Change one: the phone goes in another room

The single highest-return change, and the reason is not the notifications.

RDNE Stock project / Pexels

Silencing a phone removes the interruption from outside. It does not remove the interruption from inside, which is the decision to check it. That decision is available to you every few minutes, and each time you make it, you make it again a minute later.

Distance removes the decision. If the phone is in another room, checking it is a deliberate act with a physical cost, and the background negotiation stops.

The same logic applies to everything else that pings. Close the mail client rather than muting it. Quit the chat app rather than setting yourself to away. An application that is open is a decision that is available.

Change two: decide the task before you sit down

A large share of a work block is spent choosing what to do in it, and choosing is not free.

Decide the night before, or at the end of the previous block. Write down the specific next action, not the project. “Finish the report” is a project and will be met with twenty minutes of tidying. “Write the results section” is a task you can start.

Kaboompics / Pexels

The test for whether something is a task: can you picture the first thirty seconds of doing it? If not, it is still a project, and the first thing to do is break it down rather than attempt it.

Change three: fixed blocks with a known end

Focus is uncomfortable at the start and the discomfort is worst when there is no visible end to it.

A fixed block solves that. Somewhere between 45 and 90 minutes suits most people and most work. The specific number matters much less than the fact that it is decided in advance, because a decided end point means the first ten minutes are survivable.

Two rules make blocks work:

• The block ends when the timer ends, not when the work ends. Otherwise every block becomes open-ended and you stop starting them.
• The break is a real break. Not the same screen and a different tab. Stand up, leave the room, look at something further away than your desk.

Change four: leave a note when you stop

When something does interrupt you, and it will, the cheapest possible insurance takes ten seconds.

Before you turn to the interruption, write one line: where you were, and what you were about to do next. Not what you did, what comes next.

That note removes most of the cost of returning, because the expensive part of resuming is not remembering the task, it is rebuilding the state you had in your head. The line hands it back to you.

The same trick works at the end of a day. Stopping mid-sentence with a note about the next sentence is a far better start to tomorrow than stopping at a clean boundary with nothing written down.

The two kinds of interruption, and why the fix differs

Not all interruptions are the same thing, and treating them as one problem is why most advice fails on at least half of them.

External interruptions come from outside: a message, a colleague, a phone call, a notification. They are visible, they are somebody else’s action, and they are the ones every article addresses. They are also the smaller half of the problem for most people.

Self-interruptions come from you. Nothing pinged. You simply stopped, and opened something else. These are harder to notice because there is no external event to blame, and they follow a pattern: they cluster at exactly the moment the work becomes difficult.

That timing is the useful clue. If you catch yourself opening a new tab, the question worth asking is not “why am I so distracted” but “what did I just hit”. Almost always the answer is a specific difficulty: a sentence you do not know how to write, a decision you have not made, a piece of information you do not have.

The fix is different for each. External interruptions are solved by distance and closed applications. Self-interruptions are solved by naming the difficulty. Write the obstacle down in one line, and either solve it or decide to work around it. The urge to leave usually disappears once the thing you were leaving is stated.

Building a day that can hold a block

One block is a technique. Getting one most days is an arrangement, and the arrangement has three parts.

Put the block where your attention actually is. Most people have one part of the day where difficult work is genuinely easier, and it is not always the morning. Notice yours over a week and then defend that slot rather than the slot productivity advice recommends.

Put the shallow work somewhere on purpose. Email, messages, admin and errands are real work and they do not need your best attention. Giving them a defined slot is what stops them filling the gaps between everything else. The aim is not to do less of them, it is to stop them being the default activity whenever something gets hard.

Leave the first fifteen minutes alone. Opening messages first thing hands the shape of your day to whoever wrote to you overnight. It also spends your best attention on other people’s priorities before you have looked at your own.

A realistic version for someone with a job: one protected block, most days, at the time that suits you, with messages checked at two or three fixed points rather than continuously. That is achievable and it is most of the available benefit. A day of uninterrupted deep work is a fantasy for almost everyone with colleagues.

What to do when focus will not come

Some days the block starts and nothing happens. Three things worth trying, in order, before concluding that the day is a write-off.

Shrink the unit. If forty-five minutes feels impossible, do ten. The purpose is not the ten minutes of output, it is that starting is the hard part and a small enough unit gets past it. Momentum is far easier to continue than to create.

Change the input. If the work is writing, try talking it out loud. If it is reading, try writing a summary as you go. A task that has stalled in one mode often moves in another, and the switch is cheaper than waiting.

Check the obvious physical things. Poor sleep, no food, no water and no daylight all degrade attention, and all four are easier to fix than a motivational problem. If you have slept five hours, the honest diagnosis is not that your focus habits need work.

If none of that shifts it, the useful move is to spend the block on shallow work you were going to have to do anyway, and to protect tomorrow’s block instead. A day spent failing to start is worse than a day spent doing the admin.

What does not work

Worth saying, because these consume most of the attention given to this subject.

Apps that block other apps. They work for the exact period during which you also had enough resolve to install and configure them. The physical version, the phone in another room, does not depend on your resolve at the moment of temptation.

Trying to multitask better. For anything involving language or decisions, there is no simultaneous version. There is fast switching, and each switch has a fee.

Ambient productivity content. Reading about focus is a way of thinking about work that feels like working. It is also, more or less, what you are doing right now, so treat this article as the last one and go and do a block.

Waiting to feel ready. The willingness to start almost always follows starting rather than preceding it.

Measuring whether any of this is working

Because the output stays roughly constant whether or not you are interrupted, the usual measure of a productivity change is useless here. You cannot tell by looking at what you finished.

Two things are worth tracking instead, and both take seconds a day.

How you feel at the end. A single number out of ten for how depleted you are, written down. Over a fortnight it will show a pattern, and the pattern is what the interruption research says should move.

How many blocks you started. Not completed, started. Starting is the behaviour under your control, and it is the one that predicts everything downstream. Counting completions punishes the days where something genuinely came up; counting starts does not.

What not to track: hours logged, tasks ticked, or anything an application produces automatically. Those measure activity, and activity was never the problem. The whole finding of the study at the top of this page is that activity can look identical while the cost underneath changes completely.

The honest limits

Three, because this is a field with more confident claims than evidence.

One experiment is one experiment. The study above is a controlled laboratory task, not your job. It is good evidence that interruption is paid for in effort rather than only in time; it is not a law.

Individual differences are real, and the paper says so. Mark and colleagues reported that personality measures predicted how disruptive interruptions were. Some people genuinely suffer more from the same interruption load.

Some interruption is the job. If your work is responding to people, the answer is not to eliminate interruption. It is to have some part of the week where it is not the arrangement.

Sleep is upstream of all of it

One thing outranks every technique on this page, and it is not a focus technique.

Attention is one of the first things to degrade when sleep is short, and it degrades before you notice, which is the part that makes it dangerous as a diagnosis. A tired person does not usually feel tired at their desk. They feel normal and find the work harder, and then look for a productivity explanation for something that has a physiological one.

The practical consequence is a rule of thumb worth applying before you change anything else: if you have slept badly for several nights, do not conclude that your focus system is broken. Fix the sleep first and re-run the experiment. Changing your working method on the basis of a week of five-hour nights will teach you the wrong lesson about the method.

The same goes for the other physical inputs. Food, water and daylight are all cheap to fix and all affect attention more than any application does. None of them are interesting, which is precisely why they are skipped in favour of a new system.

The mistakes worth avoiding

• Muting instead of removing. The decision to check is the cost, and muting leaves it in place.
• Starting without a decided task. The first twenty minutes go to choosing, every time.
• Open-ended sessions. Without an end point, the beginning is harder and you start fewer of them.
• Treating a normal output day as evidence nothing is wrong. The compensation mechanism hides the cost until it accumulates.
• Optimising the tool instead of the room. A new app is a smaller change than a closed door.
• Reading more about focus. The literature is not the bottleneck.

The environment does more than the technique

Almost all attention advice is about what you do. A surprising amount of the result is decided by where you are before you do anything.

A room where the work happens. Not necessarily a study. A specific chair, a specific desk, a specific corner of a library. The point is association: a place used only for one kind of work starts putting you into that kind of work faster.

Nothing else on the surface. A desk with four other unfinished things on it is a desk that offers four alternatives to the thing you sat down for. Clearing it takes a minute and removes a set of decisions.

Noise you have chosen. Silence works for most people and for most demanding work. Where silence is not available, consistent noise beats variable noise, and anything with words in it competes directly with reading and writing.

A visible clock, and nothing else that updates. Knowing how long is left helps. A screen that changes on its own, anywhere in your eye line, does not.

None of this is exotic and all of it is cheaper than any tool. The reason environment beats technique is that it works without you having to remember to apply it, which is the same reason the phone in another room beats the phone on silent.

The bottom line

The famous number is wrong, and the corrected version is more useful: interruption does not necessarily make you slower, it makes the same work cost more, and it charges the difference to stress and effort rather than to the clock.

Which is why the fixes are physical rather than motivational. Phone in another room. Task decided in advance. A block with an end. A note when you stop.