Almost everything written about concentration assumes a room.
A door that shuts, hours that are yours, a household that understands you're working. It's good advice for people in that position, and it's the position of a fairly small number of them.
For everyone else — open-plan offices, small flats, small children, shared houses, shift patterns that overlap with other people's evenings — the useful adaptations are different, and mostly they're not about noise.
Noise is not the main problem
This is the thing worth getting right first, because it redirects the effort.
Steady noise is manageable. A busy café, a washing machine, traffic — most people can work through all of it, and plenty work better with it than in silence.
What breaks concentration is interruption, and specifically the possibility of interruption. Being asked something takes thirty seconds and costs fifteen minutes. But the larger cost is that when interruption is possible at any moment, you never go deep enough to lose it, because part of your attention stays allocated to the door.
Which is why headphones only half work. They remove the noise and not the interruptibility, and the second one is doing most of the damage.
So the useful question isn't how to make it quieter. It's how to make thirty minutes genuinely uninterruptible, which is usually a social problem rather than an acoustic one.
That reframing matters because it changes what you'd spend money and effort on. Better headphones don't buy uninterruptibility. A conversation with the people you live with does, and it's free and considerably more uncomfortable, which is presumably why the headphones get bought first.
Buying uninterruptibility
Four ways, roughly in order of how well they work.
Negotiate a specific window with the people involved, small enough that they'll agree. Not I need to work in the evenings but nobody talks to me between eight and half past nine on Tuesdays and Thursdays. Specific, bounded, and returnable — the boundedness is what makes it acceptable to everyone else, and open-ended requests get eroded within a fortnight.
Use the hours nobody wants. Before anyone's up, after everyone's down, the hour before your shift. Genuinely uninterrupted time is available in most households; it's just at inconvenient times, and the inconvenient times are the price.
Leave. A library, a café, a car in a car park. Physical absence is far more effective than any arrangement made inside the house, because you cannot be casually asked something from a different building.
Make the signal visible. Whatever it is — headphones on, a closed door, a note — it needs to be a thing others can see without asking. Half the interruptions in a shared space are people checking whether they can interrupt.
Design the work for the conditions
The other half is choosing what to do with the time you've got, and this matters more than most people expect.
Work has different tolerances for fragmentation. Editing, admin, correspondence, code review, exercises from a book — these survive being interrupted, because the state you're holding is small and rebuilds quickly. Design, difficult writing, debugging, anything where you're holding a whole structure in your head — these don't, and attempting them in an interruptible half hour produces frustration and little else.
So match them. Put the fragile work in whatever genuinely protected time exists, even if it's only two ninety-minute blocks a week. Put the robust work everywhere else. A lot of people spend their protected hours on email and then attempt the difficult thing at the kitchen table with a four-year-old present, and conclude they can't concentrate.
And write the line at the end of every session, because fragmented work is exactly where the reconstruction cost bites hardest. If your next block is Thursday and it's Tuesday now, a note saying where you stopped converts Thursday's first fifteen minutes from archaeology into work.
There's a related trick worth having: stop mid-sentence, or mid-task, rather than at a tidy boundary. Finishing a section and closing the laptop feels satisfying and makes Thursday harder, because Thursday now starts with deciding what comes next. Stopping with something obviously unfinished gives the next session a running start, which matters far more when sessions are scarce.
Where this gets glib
Two honest limits.
Some situations don't have a solution. A person in a one-bedroom flat with a newborn and no family nearby does not have two protected ninety-minute blocks a week, and no amount of negotiating produces them. Advice that implies otherwise is advice from someone who hasn't been there. The genuine answer for that period is to park most things and accept that the number of live goals is one or zero.
And every option above has a cost someone pays. Leaving the house on a Saturday morning means someone else does the childcare. An agreed evening window is time not spent with the person who agreed to it. Those are real trades, and the writing on deep work tends to present the protected hours as though they were free.
The version that survives is the one everyone has actually agreed to, which usually means smaller than you wanted.
Two hours a week
He's writing a technical book with two children under six and a full-time job. The advice he keeps reading assumes an office.
What he tries first is evenings after bedtime — theoretically two hours, practically interrupted about half the time, and the interrupted half is spent on nothing because the work needs continuity.
What works instead has three parts. Tuesday and Thursday, half seven to nine, agreed with his partner and reciprocated with two evenings of his own where she gets the same. Those are the protected blocks, and they're for drafting, which is the fragile work.
Saturday mornings he leaves the house and goes to the library for two hours, which is unambiguous in a way that being at home never is.
And everything else — checking references, tidying code samples, responding to reader emails — happens in fragments, on a phone, in ten-minute pieces that would be useless for drafting.
Roughly five protected hours a week, plus fragments. Not the clear afternoons the advice assumes. Four chapters in six months, which the clear afternoons had not produced in the previous year, because they never happened.
What gets counted
Two ninety-minute blocks a week doesn't look impressive against advice built on daily deep work, and the counter doesn't care.
Days since you last touched it is the measure, and Tuesday-and-Thursday keeps it under five permanently. A goal at four days quiet is alive, whatever the total hours are, and the total hours were never the thing that determined whether it survived.
For work in fragments, the notes matter more than usual. They're what let a Thursday evening pick up from a Tuesday, and without them, a schedule of two blocks a week means half of each block spent remembering.
Related: Ninety minutes is the unit · One line is enough · Seasons: when it's correct to have no goals at all
Share: