Some weeks die by Tuesday. A client moves a deadline up, the laptop decides to spend a morning on updates, a kid stays home from school, and by Wednesday the plan I made on Sunday is a fiction. I used to try to salvage the plan anyway, which meant working evenings to honor a schedule that no longer described my life. Now I run a reset instead. It takes 20 minutes, it runs the same way every time, and it gives me back the rest of the week.
The twenty-minute reset
I do it at a clear desk with a timer running, and I use paper first because screens lead me back into the inbox.
Step one: dump. Everything still on my head goes onto one list, unsorted. What I owe, what I promised, what I’m avoiding, what I forgot to write down last week. Seven to twelve items is normal.
Step two: mark the hard deadline items. Anything that must happen before Friday at 5pm gets circled. In a wrecked week that’s usually three or four items, and seeing that number is most of the therapy.
Step three: move the rest, out loud. “Out loud” means sending the message: the reschedule, the new date, the polite admission that this week has beaten me. Nothing gets silently pushed to a vague “later.” The average reset sends about five messages.
Step four: rebuild the calendar for the days that are left, and I’ll come back to the capacity rule below.
What I cut first
The test I use: would I agree to this today if it were offered fresh? If the honest answer is no, it gets moved. Plenty of what’s on a broken week’s list were yeses given in a world that has since changed, and a commitment made under old conditions is allowed to expire.
The research on implementation intentions backs the structure here: plans that fix the when and the where survive interruption better than plans that just name the goal. A reset is me rebuilding those when-and-where slots around the week I actually have.
The half-capacity rule
The remaining days get planned at 50% of what I think I can do. This is the least intuitive step and the one that made the whole thing work.
A week that has already gone sideways has proved that my estimates are fiction. So I book one deep block, 9 to 11, for the single most important thing, and leave the afternoons as catch-up or meetings. If the estimate turns out to be right and I finish early, the extra hour is a gift. If it turns out to be wrong again, the damage lands in empty time instead of on top of more promises.
I also check the working-hours settings in Google Calendar before I rebook anything, because half my reset cases start with meetings that wandered into hours I had marked as unavailable.
The Friday note
Three lines, in a plain text file I keep called resets.txt: what slipped, why, what I change. Over 2025 I logged 19 resets, and 11 of them traced to the same cause. I had booked five deep-work blocks in a week that already held two meetings a day. That pattern was invisible week to week and obvious in the file.
The plan is disposable. The system for replacing it is what I protect.