I Stopped Defending My Focus and Started Rebuilding It. The Rebuild Is the Skill.

For a year I treated focus as a thing you defend — a block of time you build walls around and hope nobody knocks over. It never survived. Every method that assumes an undisturbed morning failed by ten o’clock, and I spent more energy being angry about the interruption than doing the work. What actually changed things was giving up on defending and learning to rebuild, because the rebuild is the part you can control.

There is a reason “just get back to it” does not work. Attention does not snap back; it leaves a residue. When I get pulled away mid-sentence, a fragment of my brain keeps chewing on the thing I was about to type, and for the next while I am technically at my desk but not actually back. The research name for it is attention residue, and the practical point is that a five-minute interruption costs me closer to twenty minutes of real focus if I do nothing to recover.

So I built a small re-entry routine and it is the highest-leverage twelve minutes in my day. First, I do not reopen the thing I was working on. I write one line on a scrap — where I was, the next three words I meant to type — so the open loop closes. A daily plain-text log can hold that re-entry note if loose scraps are hard to find. Then I stand up, physically, for ninety seconds; moving resets the body in a way that telling myself to refocus never did. Then I set a timer for twelve minutes and do nothing but the one open loop from the scrap. Twelve minutes of real work beats an hour of pretending.

The other half of rebuilding is triage, and it is the part I resisted longest. When I come back, I ask one question: is this fragment worth finishing now, or worth rescheduling whole? Most of the time the honest answer is the second, and fighting that is what turns a broken block into a ruined afternoon. Optimisation assumes you get your schedule back. Triage assumes you will not, and works anyway.

I have come to think of the ninety seconds after an interruption as the actual lever, not the hours before it. What I do in that window — close the loop on paper, move, pick one thread — determines whether I recover in minutes or in an afternoon. Routine is what you had planned; recovery is what you actually do. One of those two things is under my control, and it is not the routine.

I tracked it for two weeks out of embarrassment more than science. On an average day I was interrupted twelve times — messages, questions, a build breaking, a meeting running long. Twelve times a day, five days a week, is sixty recoveries a week, and each one either costs me five minutes or twenty depending on what I do with it. That is the real budget. Nobody who writes about deep work talks about sixty recoveries a week, which is exactly why their advice did not hold.

The thing I finally let go of was the fantasy of the undisturbed morning. I no longer block three sacred hours, because a three-hour block that gets broken at minute twelve is not three hours — it is a broken promise I then resent. Instead I protect smaller, uglier, more numerous blocks and spend the energy I used to spend on walls on the re-entry instead. It is less romantic and it actually holds.

The transferable bit, if there is one: plan for the interruption, not against it. Assume the block will break, decide in advance what you will do in the ninety seconds after it does, and treat the twelve minutes after that as the work, not the recovery from the work. Defending a block is a wish. Rebuilding it is a skill, and skills are the only part of this you can practice.