Deep work is a schedule problem before it is a discipline problem

The standard advice for concentration is personal: reduce distraction, build discipline, meditate. The larger determinant is usually structural, and it is not within the individual's control.

The interruption research

Studies observing knowledge workers have found interruptions occurring every few minutes on average, from both external sources and self-interruption.

The frequently cited figure for recovery after an interruption — around twenty-three minutes to return to the original task — comes from observational work and should be read as an average across a wide distribution rather than a precise constant.

The direction is robust even where the number is soft: resumption is not instantaneous, and the cost is not the duration of the interruption.

Attention residue

The more useful concept.

Research on task switching describes a residue effect: after switching, part of attention remains on the previous task, particularly when it was left incomplete.

Which means the cost of switching is not only the reorientation time but a degraded quality of attention on the new task for a period afterward.

Practically, a day composed of many short blocks is not equivalent to the same total hours in a few long ones. The fragmented version produces less, and the person doing it correctly perceives that they worked hard and achieved little.

Why the calendar is the actual problem

A day with meetings at ten, twelve and three contains no usable block, despite containing several free hours.

A one-hour gap before a meeting is not an hour of deep work. It is an hour of anticipating a meeting, which is a different and much less productive state.

Which means the fix is not personal discipline but calendar architecture: consolidating meetings into defined periods so that the remainder is contiguous.

Teams that adopt meeting-free days or shared focus blocks report substantial improvement, and the individual could not have achieved it alone.

What to do individually

Where the structure cannot be changed, several things still help.

Defend one block, not the whole day. Ninety minutes of genuinely protected time is worth more than a vague intention to focus throughout.

Schedule it as a meeting, because an appointment with yourself is respected by calendar software and, in most organisations, by colleagues.

Put it where your attention actually is. Circadian variation in alertness is real and individual. Most people have a period of better concentration, frequently in the morning, and spending it on email is a substantial waste.

Leave a task incomplete deliberately. Stopping mid-sentence or mid-problem makes resumption dramatically easier, because the re-entry point is obvious.

This is a genuinely underused technique and it costs nothing.

Capture, do not chase. When an unrelated thought arrives — and it will — write it down rather than acting on it. The interruption is then two seconds rather than twenty minutes.

The notification question

Worth being specific about.

Turning off notifications helps and it is not sufficient, because a substantial share of interruptions are self-generated. People check their own phones without any prompt.

The mechanism is habitual and cue-driven, which means the effective intervention is physical distance rather than restraint. A phone in another room is checked far less than a phone face-down on the desk.

What deep work is not

Worth stating, because the concept has been oversold.

Not all valuable work is deep work. Coordination, responsiveness, review and conversation are real work with real value, and a role composed entirely of solitary concentration is unusual.

The argument is not that shallow work is worthless but that most schedules have drifted so far toward it that the deep component has been squeezed to nothing without anyone deciding to do that.

The realistic target for most people is not a day of uninterrupted concentration. It is one protected block, most days, which is achievable and which is more than most currently have.

Protecting the block from yourself

A calendar entry is necessary and not sufficient, because the person who arrives at the block is not the one who scheduled it.

What helps is deciding the specific task the night before, leaving the document open at the point of resumption, and refusing to check messages first.

That last one matters more than it sounds. A block that begins with the inbox has already been spent, because the contents of the inbox now occupy the working memory the block was for.