Plan low-energy afternoons realistically

The 15:00 architecture slot that never worked

For about a year my weekly plan had the same quiet lie in it. Mornings filled up with code reviews, client replies and a standup, so the hard thinking work got pushed to the afternoon. I would write "design sync model for the app, 15:00 to 17:00" on a Tuesday with total confidence, and at 15:20 I would be rereading the same three files, then opening chat to feel useful, then promising myself an hour after dinner.

The plan was fine on paper. What it ignored was that I am a noticeably different worker at three in the afternoon than at nine in the morning. I knew that in a vague way. I had just never planned around it, because planning around it felt like admitting something.

This is not an argument that everyone fades after lunch. Some people do their best work at four. It is an argument that energy is a schedule input, like meetings or a school run, and that a plan which ignores it will fail in the same place every week.

Photograph

Two weeks on an index card

What finally changed my planning was a very low-tech experiment. For two working weeks I kept an index card next to the keyboard and, at lunch and again at shutdown, marked the half-day just finished as high, medium or low. High meant I could hold a hard problem in my head. Medium meant I could review, edit or follow an outline. Low meant I could do things, but not think about them much.

Twenty half-days later the pattern was clearer than I expected:

• Mornings were high on eight days out of ten. The two exceptions followed bad nights.
• Afternoons after a call-heavy morning were low every single time.
• Friday afternoons were low regardless of anything else.
• Afternoons when I had gone outside at lunch were usually medium, not low. Not high, but useful.

None of that is science. It is one person's card. But it was my card, and it beat the plan I had been writing from optimism. I've found the exercise worth repeating every year or so, because the pattern shifts when work or home routines change.

If you try it, record what you observe, not what you wish. The temptation is to mark a dull afternoon as medium because low feels like a confession. It is just data.

Match the task to the gear you are in

Once you know the shape, the planning rule is simple: high windows own the Primary, the one piece of work that needs a loaded brain. Low windows get work that is genuinely useful but does not need much composition. Medium sits in between.

The other half of the rule is easy to miss: do not spend high windows on low work. If your sharpest two hours go to the inbox, the afternoon cannot rescue the day. Protecting two uninterrupted hours is about defending exactly that window.

Build a slump menu before you need it

The real problem at 14:30 is not low energy. It is low energy plus a decision. Tired me, asked "what should I do now?", picks chat every time. So I keep a written list of honest shallow tasks, refreshed monthly, pinned where I can see it.

Mine currently has things like:

• Clear the review queue down to anything older than a day
• Send status updates on the two client projects, with next and blocked
• Process receipts and log the week's hours
• Bump one dependency and run its tests
• Write the staging notes for yesterday's merge
• Tidy the issue tracker for one of my own apps
• Draft, but do not send, any awkward email I have been avoiding

Every item is useful, finishable in thirty to forty-five minutes, and does not need me at my best. That last point matters. A menu full of "small" tasks that secretly require deep thought just moves the problem.

The same Thursday, planned twice

Assumptions: standup at 10:00, sharp from 08:30 to 12:00, fading after 13:30, and a hard stop at 15:30 for a school run.

The plan I used to write. Primary is "finish the checkout tax refactor", scheduled 13:30 to 15:30 because the morning was full of reviews. At 14:00 you reread the same files. At 14:40 you open chat. At 15:20 you feel behind, promise an evening session, and leave for the school run with the refactor in pieces and no note about where you were. The evening session is shallow. Friday morning starts tired.

The plan I write now. 08:30 to 10:00 owns the refactor, with a finish line of "pull request open, or the failing assertion pasted in the ticket". Standup at 10:00. 10:30 to 12:00 is reviews, which are medium work and fit a late morning fine. After lunch, 13:30 to 14:30 is review leftovers and the client status update. 14:30 to 15:00 is one item from the slump menu. 15:00 is the shutdown note, naming tomorrow morning as the next Primary slot. Out the door at 15:30 without guilt.

Same hours. Same person. The difference is that the second plan finishes something reviewable before the fade, while the first spends the fade pretending to be morning. Planning eight hours into eight hours fails partly for this reason: it treats every hour as interchangeable.

When the afternoon is the only window you get

Not everyone gets to choose. Caregiving, time zones or a client who only takes calls in the morning can make the afternoon the only time left for deep work. If that is your situation, do not pretend the fade is not there. Shrink the deep work to fit it.

Shorter blocks with sharper finish lines beat long empty attempts. A twenty-five or fifty-minute block with a tiny target ("the validation function written and one test passing") gets done on a flat afternoon where a heroic ninety minutes would not. When a 25-minute Pomodoro is too short covers choosing the length; in the afternoon, I go shorter.

If the household schedule breaks the afternoon in two, plan the pieces separately. A school run at 15:30 often drains focus from about 14:00, so put one light task in that gap instead of a refactor. A remote workday with caregiving interrupts goes into that shape in more detail.

Shallow mode, said out loud

On the afternoons when I sit down and the Primary is plainly not going to happen, I say "shallow mode until stop time" to myself, then pick from the menu. It sounds a bit silly. It works because guilt fills an unnamed gap with a story about failing, and naming the mode replaces that story with a plan.

Then two rules. Set a hard stop, and do not make up the depth at nine in the evening. And write on the shutdown note when tomorrow's Primary will happen: morning or afternoon. That one line stops you from planning a heroic morning on a day you already know will fragment.

Batch the shallow work into one window with a timer, so the slump has a job and a container. When batching helps is a good companion here. Without the container, a shallow afternoon becomes an afternoon of snacking on chat.

Traps worth naming

The universal morning myth. Not everyone peaks early. Use your own card for two weeks before rewriting your week.

Meeting stacking in the only sharp window. If calls land on your best morning hours, defend at least one sharp hour somewhere else, even if that means a split day (structuring a remote workday).

Fake deep work. The editor open for two hours with nothing to show feeds guilt. An honest admin batch at least feeds closure.

Telling nobody. If clients and colleagues know you do deep work in named windows and send updates on schedule, they rarely care which hour produced it. Demos and hard conversations go in sharp hours. Ticket hygiene goes in the valleys.

I still sometimes write "architecture, 15:00" out of habit. Then I cross it out and move it to tomorrow at 08:30, and the afternoon gets the receipts and the reviews it was always going to get. Planning the afternoon you actually have is less impressive on paper and finishes a lot more by Friday.