How to get an hour of real focus in a distracting job

Books about focus tend to assume you can disappear for four pristine hours into a monastery of deep thought. Most people have jobs with colleagues, chat channels, and managers who occasionally need things. The realistic target is smaller and, it turns out, sufficient: one genuinely protected hour of focused work, most days. One good hour, done consistently, moves almost everything that matters — and unlike the four-hour fantasy, it survives contact with a normal job.

This is the long version of how to get that hour: what interruptions actually cost according to the research, how to protect a window, what to do about notifications and the phone, and how to recover when the hour gets invaded anyway.

What interruptions actually cost

The best-known researcher in this area is Gloria Mark, Chancellor's Professor of Informatics at UC Irvine, who has spent two decades measuring what digital work does to attention. Her findings are bleak in an instructive way. In an interview with the American Psychological Association, she reports that average attention on a screen before switching has fallen from about two and a half minutes in 2004 to roughly 47 seconds today.

Her lab work shows what those switches cost. In the much-cited study "The Cost of Interrupted Work: More Speed and Stress," Mark and colleagues found something subtler than the popular retelling: interrupted people actually completed tasks faster, with no drop in quality — but reported significantly more stress, frustration, time pressure, and effort. We compensate for interruptions by sprinting, and the sprint has a physiological price. (The widely quoted claim that refocusing takes exactly "23 minutes" is hard to pin to any published paper; the documented costs are the stress and the constant switching itself.)

The scale of the problem is broad: in Microsoft's 2023 Work Trend Index, 68 percent of 31,000 surveyed workers said they don't get enough uninterrupted focus time during the workday. In other words, the deck is stacked against focus by default. Getting an hour requires deliberately unstacking it.

Why one hour beats chasing four

The all-or-nothing instinct — "if I can't get a long, pristine session, deep work doesn't count today" — is how busy people end up doing no deep work at all. One hour is short enough to defend on a bad day and long enough to make real progress on something hard. Done most days, it compounds into a startling amount of finished work over a month.

There's also a quality argument. Demanding cognitive work has a daily ceiling — practitioners like Cal Newport, who built his career writing about deep work, put it at roughly three to four hours even under ideal conditions. The second and third "deep" hours of a tired afternoon are often theater: sitting at the desk looking serious while the brain idles. One hour of real attention beats three hours of performance.

Treat the hour as the floor, not the ceiling. On days when the calendar allows a second window, take it. But the system's promise is the floor: one defended hour, almost regardless of what the day does.

What counts as deep work for the hour? Anything that needs sustained reasoning and produces something: writing, analysis, design, code, preparing a difficult conversation. What doesn't: email, status updates, anything that could be done adequately while half-listening to a call. The hour is too scarce to spend on work that doesn't need it.

Protect a daily window

The most effective single move is claiming the same window every day, ideally the first hour of work, before the organization fully wakes up. Early mornings are naturally protected: fewer messages, fewer meetings, fewer people expecting an instant reply. A consistent window also trains colleagues without a single awkward conversation — after two weeks of "she's heads-down until ten," people simply route around it.

Three details make the window hold:

• Put it on the calendar as a real block, visible to colleagues, so the slot defends itself against meeting invites. This works best as part of a broader approach to time-blocking.
• Start the focused work before opening email or chat. Once the inbox has been seen, attention is already snagged on other people's priorities, and the hour quietly becomes triage.
• Pick the window where the world demands least of you. It doesn't have to be morning. A brain that comes alive at 4 p.m. should protect 4 p.m. The principle is putting the hardest work where interruption pressure is lowest, rather than hoping focus appears spontaneously mid-afternoon. It won't.

Triage your notifications

Nobody can focus while a device is actively trying to interrupt them, and most devices are, constantly, by design. Most notifications serve the app's engagement metrics, not the recipient. The fix is a one-time triage, sorting every notification source into three levels:

The triage takes twenty minutes, once. During the focus hour itself, go one step further: do-not-disturb on, chat and email apps fully closed rather than minimized — a closed app shows no badge — and irrelevant browser tabs shut, because an open tab is an open invitation.

The batched middle tier deserves emphasis, because it's where most knowledge-work interruptions originate. Checking chat and email at scheduled points rather than on arrival converts dozens of daily interruptions into two or three planned sessions, and the evidence on email specifically supports it — a University of British Columbia experiment found that limiting email checks to three times a day measurably reduced daily stress. The full email routine is covered in how to get your inbox under control.

The fear is missing something important. In practice, anything truly important arrives through a channel that is still open (a call), and everything else can wait sixty minutes. The constant low hum of "something might be happening" is itself a cost — and it stops within days of going quiet.

Tame the phone without deleting everything

The phone deserves its own section because it undermines focus even when it's silent. In a 2017 study of nearly 800 participants at the University of Texas at Austin, Adrian Ward and colleagues found that the mere presence of a smartphone reduced measurable cognitive capacity — performance on concentration tests was best when the phone sat in another room, worse in a pocket or bag, and worst on the desk. The effect held even when the phone was powered off. Part of the brain, Ward argues, spends resources actively not-using the visible phone.

So the headline rule is physical: during the focus hour, the phone goes in a drawer or another room. Not face down on the desk — out of sight. The thirty seconds it takes to go fetch it is usually long enough to realize it wasn't needed.

For the rest of the day, the goal is making the phone boring rather than forbidden. Cold-turkey approaches fail because the phone is also the map, the camera, and the family channel; deprivation that ignores the genuinely useful parts collapses within days. What works is friction, applied selectively:

• Bury the time-sinks. Move the apps that eat hours off the home screen into a folder on the last page, and log out of the worst ones so each use requires a deliberate sign-in. The reflexive open — the one that happens without any decision — dies when reaching the app takes three swipes and a password.
• Use app timers as tripwires, not walls. A "you've been here 20 minutes" nudge breaks the trance even if you tap through it.
• Make the home screen dull. Keep only tools on page one: maps, camera, calendar, notes, messages. Grayscale mode, used during work hours, removes more pull than seems plausible — color is part of what makes apps feel rewarding.
• Set two default phone-free zones — the dinner table and the bedside — so the first and last act of the day isn't a screen.

None of this is heroic, and none of it requires giving up anything actually valued. It just shifts the default from "the phone interrupts unless restrained" to "the phone serves when summoned."

One task, decided the night before

A protected, silent hour can still be wasted on deciding what to do with it. Fifteen minutes of deliberation eats a quarter of the hour and most of its freshness.

The fix: at the end of each workday, pick the one thing that gets tomorrow's focus hour, write it down where it will be seen, and leave the relevant document open. One thing — not a list. This is an implementation intention in miniature: the start decision gets made at a calm moment instead of in the morning fog, when everything feels equally urgent and the inbox beckons. When the hour arrives, there is nothing to decide, only a task waiting with the door already open.

Size the task to the hour. "Finish the report" sets the hour up to fail; "draft the findings section" sets it up to succeed. An hour that ends with something visibly completed builds the habit loop — protected time produces finished work produces the will to protect tomorrow's hour — while an hour that ends mid-everything just feels like more unfinished business.

When you get interrupted anyway

Some days the hour gets invaded regardless: a genuine emergency, an unmovable meeting, a colleague who really does need help now. A focus system that shatters on contact with reality is useless, so build in the recovery moves.

Defend gently first. For non-urgent drop-ins during the focus hour, a warm, specific deferral — "Can I come find you in an hour?" — works far better than a vague brush-off. Most things can wait an hour, and people respect a concrete commitment. (This is a small instance of a larger skill, covered in how to say no at work without burning bridges.)

If interruption is unavoidable, break at a boundary. Mark's research distinguishes interruptions at natural break points — the end of a paragraph, a finished sub-task — from interruptions mid-thought, and the mid-thought ones cost far more re-orientation. Given any choice at all, finish the sentence before turning around.

When the hour is lost, find a smaller one. Thirty minutes after lunch beats zero. The goal is consistency over perfection: enough good hours across a week that the important work moves, not an unbroken streak. Some days the job simply eats the focus, and that's fine — the system's job is to make those days the exception instead of the default, and to make sure a bad Tuesday never argues you out of protecting Wednesday's hour.

Frequently asked questions

How long does it take to refocus after an interruption?

You may have seen "23 minutes" quoted as gospel. That exact figure comes from interviews rather than a published table, so treat it as folklore-adjacent. What Gloria Mark's published field and lab studies do show clearly: interrupted workers pay measurable costs in stress, frustration, and time pressure, and screen attention now averages well under a minute between switches. The precise minute count matters less than the direction: interruptions are expensive.

Does music help or hurt focus?

It depends on the task and the music. For verbal work like writing or reading, lyrics compete for the same language circuitry and tend to hurt. For routine or well-practiced tasks, familiar instrumental music can help by masking office noise and flattening distractions. A practical rule: the harder the thinking, the quieter the soundtrack — and silence or steady ambient noise beats anything novel.

Is multitasking ever fine?

For genuinely demanding cognitive work, no — what feels like multitasking is rapid switching, and each switch carries a cost in errors and re-orientation time. Pairing a mindless physical task with a podcast is fine; pairing two tasks that both need working memory is not. If both tasks involve language, screens, or decisions, do them one at a time.

Do focus apps and website blockers work?

They help as scaffolding, not as a cure. Blockers reduce the cost of a weak moment by adding friction at exactly the right time, which is the same principle as putting the phone in another room. But they work best alongside a defended calendar block and silenced notifications. A blocker fighting against an otherwise interruption-friendly setup loses eventually.