# How to Time-Block a Deep-Work Day That Survives Meetings

> Plan a deep-work day with block budgets, buffers tuned to your own log, a two-minute re-blocking routine for broken days, and templates for five roles.

- Canonical page: https://learnhubplay.nl/blog/time-blocking-deep-work-day
- Published: 2026-10-11
- By: LearnHubPlay, an independent software studio in Amsterdam
- Topic: Focus & Deep Work
- Tags: Time management, Deep work, Planning, Meetings, Templates

A time-blocked day often breaks the same way. At 08:30 the calendar is a clean grid of focused hours. By 11:00 a meeting has landed in the middle of it, the first block has run twenty minutes over, and the plan describes a day that no longer exists. Plenty of people decide time blocking does not suit their job and go back to a to-do list.

The grid was not the problem. A plan built as exact promises breaks on contact; one with slack, a ranked order of sacrifice and estimates corrected by your own history bends.

By the end you will be able to plan tomorrow in ten minutes, recover a broken day, and size blocks to your track record rather than your optimism.

## Why a tidy plan runs short: the planning fallacy in numbers

Psychologists call the habit of underestimating your own tasks the *planning fallacy*, a term Kahneman and Tversky introduced in 1979. In Study 1 of Buehler, Griffin and Ross (1994), 37 psychology students at the University of Waterloo were phoned while writing honours theses and asked when they would submit. Each gave a best guess, a best case and a worst case:

| Prediction | Average predicted | Average actual | Finished by then |
|---|---|---|---|
| Best guess | 33.9 days | 55.5 days | 29.7% |
| Best case | 27.4 days | 55.5 days | 10.8% |
| Worst case | 48.6 days | 55.5 days | 48.7% |

Means are for the 33 who finished; four more had not submitted two semesters later, so the real miss is larger. Even the worst case fell short, though not by a statistically significant margin.

Predicted and actual times correlated at r = .77: the ranking of tasks was sound, the level too low. In Study 4 (123 students, a one-hour computer assignment), students who linked past experience to the task predicted 7.0 days and took 7.0 on average; 60% finished by their prediction against 29.3% of controls. Case by case they were no more accurate: the average bias vanished, the scatter did not.

These studies measured calendar days, not focused minutes, and used students. We borrow the direction of the finding, not its size; until you have your own numbers, treat your best guess as a lower bound.

## Three block types and a starting budget for a workday

| Type | Job | Length | Starting budget |
|---|---|---|---|
| Deep | One named result, uninterrupted | 60–90 min | 2 blocks, about 3 h |
| Shallow batch | Chat, email, reviews and admin together | 30–60 min | 2 batches, about 1–1.5 h |
| Buffer | Empty on purpose; absorbs overruns and spillover | 30–45 min | 20–25% of working time outside meetings and breaks |

The lengths and shares are our starting defaults, not research findings. Plan fewer deep hours than feels possible, and raise the cap only when your log shows you finishing the blocks.

The buffer is the unusual one. Cal Newport [says he first described time blocking on his blog in 2013](https://calnewport.com/the-time-blocking-revolution-begins/) and covered it in *Deep Work* (2016); his [handwritten plans leave extra room next to the blocks](https://calnewport.com/deep-habits-the-importance-of-planning-every-minute-of-your-work-day/) for corrections. A buffer builds the same idea into the schedule.

## Where meetings go: Graham's maker and manager schedules

Paul Graham's July 2009 essay "Maker's Schedule, Manager's Schedule" contrasts two ways of using a day. The manager's day is cut into one-hour intervals, so a meeting costs an hour. Makers prefer units of at least half a day, and, in Graham's words, "a single meeting can blow a whole afternoon, by breaking it into two pieces each too small to do anything hard in."

Three rules follow:

- **The gap rule.** Treat any open gap under 60 minutes as shallow or buffer time. The 60 is our threshold, not Graham's; raise it if your work needs a long warm-up.
- **One cluster.** Stack meetings back to back so the day has one large gap, not three fragments.
- **Counter-offer next to something.** Reply to a proposed time with a slot adjacent to a meeting you already have.

Teams use meeting-free days for the same reason. A [2022 MIT Sloan Management Review article](https://sloanreview.mit.edu/article/the-surprising-impact-of-meeting-free-days/) on 76 large companies reported gains, based on surveys, interviews and before-and-after comparisons, not a controlled trial.

> **Warning:** Google Calendar's Focus time can decline invitations for you. On a computer, open Day or Week view, click the time, choose Focus time and tick "Automatically decline meetings." The default, "New and existing meetings", can also decline a meeting already on your calendar that overlaps the block; choose "Only new meeting invitations" to avoid that. It needs a work or school account on Business Standard or Plus, Enterprise Standard or Plus, Education, or Nonprofits. In other calendars, mark the block Busy.

## Build tomorrow's blocks in ten minutes

Do this at the end of the day. Newport spends [ten to twenty minutes each evening](https://calnewport.com/deep-habits-the-importance-of-planning-every-minute-of-your-work-day/) on the next day's plan; these steps fit ten.

1. **Enter fixed events first:** meetings, commute, school runs, hard stops.
2. **Label every gap** with the gap rule: 60 minutes or more is a deep candidate.
3. **Name the first deep block's result and first action.** "Draft sections 2 and 3 of the import design; write the section 2 heading" beats "work on import."
4. **Estimate by steps.** List three to five steps, estimate each, add them up; in three experiments, [Forsyth and Burt (2008)](https://doi.org/10.3758/MC.36.4.791) found people gave a single task less time than its subtasks combined. Multiply by your ratio from the log below, or ×1.5 until you have five entries for that task type (the thesis writers ran at about ×1.6).
5. **Place two shallow batches** away from your best hours, and **two buffers**: one straight after the meeting cluster, one near the end.
6. **Write the sacrifice order** on the plan (see re-blocking below).

A paste-ready plan:

```text
TOMORROW <date> | Fixed: <meetings, hard stops>
Deep 1 <time>: result ___ / first action ___
Deep 2 <time>: result ___ / first action ___
Shallow <time>, <time> | Buffer <time>, <time>
Sacrifice order: 1 buffer, 2 move shallow later, 3 shrink Deep 2, 4 Deep 1 last
```

### A worked example day

Hypothetical day for a developer (start, block, minutes):

```text
08:30  Check plan, write first action         15
08:45  Deep 1: import design, sections 2–3    90
10:15  Break                                  15
10:30  Deep 2: importer tests, two edge cases 90
12:00  Lunch                                  45
12:45  Batch 1: chat, email, review requests  45
13:30  Meetings: one-to-one, sprint planning  90
15:00  Buffer 1: overruns and follow-ups      30
15:30  Batch 2: reviews, admin                30
16:00  Buffer 2: whatever slipped             45
16:45  Shutdown and tomorrow's plan           15
```

That is 510 minutes: 180 deep, 105 shallow, 90 meetings, 75 buffer, 60 breaks and lunch. The buffer is about 21% of the 360 working minutes outside meetings and breaks.

### Role templates: maker, manager, on-call, student, parent

Hypothetical starting points:

| Role | Deep | Buffer | Rule to remember |
|---|---|---|---|
| Maker | 2 × 90 min before midday | 2 × 30 min | Two shallow batches of 45 min; meetings in one cluster, same half-day |
| Manager (4–6 meetings) | 1 × 60–90 min before the first meeting | 30 min between clusters | Three short batches; block open-door time like a meeting |
| On-call or support | 1 × 60 min, interruptible | 90 min total | Queue windows every 90 min; give reactive blocks a fallback task ([Newport](https://calnewport.com/deep-habits-the-importance-of-planning-every-minute-of-your-work-day/)) |
| Student | 2 × 50–60 min in gaps of 60+ min | 30 min after the last lecture | One admin batch; treat lectures as meetings |
| Parent or carer (school runs) | 1–2 × 60–90 min in school hours | 30 min ending at a hard stop | One batch before pick-up; end blocks 15 min before it |

## When the day goes sideways: a two-minute re-blocking routine

Newport's [2015 write-up of three daily plans](https://calnewport.com/deep-habits-three-recent-daily-plans/) shows the habit. The columns growing to the right of a plan are rewrites made as the day changed: a visitor during his 12:30 block shifted his 1:30 task block, which "ended up lasting until 3." A rewrite is the method working, not failing.

1. **Gray out what is gone:** the interruption and the blocks it destroyed.
2. **Count the minutes** to the next fixed event. Only that stretch can absorb the loss, because work moves later, not earlier.
3. **Spend in sacrifice order** inside that stretch: a buffer first, then move shallow work to the first batch after the fixed event, then shrink the second deep block's scope. The first deep block goes last.
4. **Write a resume note:** what is done, what is open, your first action on return. The [attention-residue guide](https://learnhubplay.nl/blog/attention-residue-task-switching) has a one-minute template.
5. **Rewrite beside the old plan**, not over it, so you can see what the day cost.
6. **Name what did not fit.** Move it to tomorrow on purpose; if it has already moved twice, split it into one-block steps or drop it and write down that you did.

On the example day: a 30-minute production alert hits at 09:45, with 60 of Deep 1's 90 minutes done. Only time before the 13:30 meetings can absorb the loss, and no buffer sits there, so shallow work goes first: Batch 1 moves after the meetings and merges with Batch 2 into one 60-minute batch at 15:30. Deep 1 finishes 10:30–11:00, Deep 2 keeps its 90 minutes at 11:00–12:30, and lunch moves to 12:30–13:15, leaving 15 spare minutes. All 180 deep minutes survive; the afternoon pays, with 60 minutes of shallow work instead of 75 and 45 minutes of buffer instead of 75.

| What happened | First move |
|---|---|
| Meeting invited into a deep block | Counter-offer an adjacent slot; if refused, write the resume note, attend, re-block |
| Morning lost entirely | Make the largest remaining gap of 60+ minutes the deep block; shrink the goal to its smallest useful slice |
| No gap of 60+ minutes all day | Call it a shallow day: continue the task that has a resume note, and put small deeper-work blocks at the edges of the day, [as Newport suggests](https://calnewport.com/deep-habits-the-importance-of-planning-every-minute-of-your-work-day/) |
| Block finished early | Do not add scope. Log the real finish time and pull the next item forward |

### Guard the first deep block hardest

Each overrun pushes everything behind it, so the first deep block of your best hours is the only one whose start you control. Mark it Busy, silence chat, and settle your alerts with the [30-minute notification audit](https://learnhubplay.nl/blog/notification-audit-mac-chrome-phone).

If you use a blocker, schedule it to match. FocusWall Pro, a Chrome extension we build, can start focus sessions on a schedule and redirect distracting tabs to a dashboard showing your target goal, with friction levels up to a Hard Mode where bypass is locked. [Why blockers get switched off, and how graduated friction helps](https://learnhubplay.nl/blog/blocking-distractions-with-friction) covers the levels. A closed chat app and Do Not Disturb work too.

## Calibrate with two weeks of planned versus actual

After each deep block and shallow batch, add one line to a plain-text log:

```text
date   | block  | task type    | planned | actual | done at block end? | note
12 Oct | Deep 1 | doc section  | 60      | 100    | yes                | ran into the break
12 Oct | Deep 2 | debug a test | 45      | 110    | no                 | 90 done + 20 estimated
```

Three rules keep it honest:

- Log the step-sum estimate you planned with, so the ratio corrects that method.
- If the block ended before the task did, the actual is a floor: minutes done plus your estimate of the remainder.
- Log the minutes the task took, not the block length, or the log just echoes your calendar.

After ten working days, divide actual by planned per task type and take the median. Say your last five "write a doc section" ratios are 1.2, 1.5, 1.7, 1.7 and 2.0: the median is 1.7. Replace the 2.0 with a 3.5, the day of the outage, and the mean jumps from 1.6 to 1.9 while the median stays 1.7. With fewer than five entries, use the median across all deep blocks.

Multiply your step-sum estimate by the ratio: with a median of 1.7, a 60-minute doc section becomes 100 minutes. Fit it into a 90-minute block by cutting scope, not by hoping.

Size the buffer the same way: for each day, add the minutes blocks ran over to the minutes of unplanned interruptions, take the median across the ten days, and round up to the next 15. That replaces the 20–25% default. Keep some buffer even when your ratios look tidy: in Study 4, linking past experience removed the average bias but not the case-by-case error.

End each day with a shutdown that doubles as planning. Newport's [2009 version](https://calnewport.com/drastically-reduce-stress-with-a-work-shutdown-ritual/) updates task lists, reviews the next two weeks of calendar and the plan, then closes the computer with "schedule shutdown, complete." Masicampo and Baumeister (2011) found in laboratory studies that unfinished goals produced intrusive thoughts and worse performance on an unrelated task, and that a specific plan removed the effect. These were one paper's lab tasks, not workplace output, and we did not check for replications, so treat the result as suggestive. Once a week, do the same at larger scale in a [weekly review](https://learnhubplay.nl/blog/weekly-review-template).

## Where to start this week

- **Today:** name the result and first action of tomorrow's first deep block.
- **Monday:** apply the gap rule, move one meeting next to another, write your sacrifice order.
- **Tuesday to Friday:** log planned and actual minutes; after week two, compute your ratios and set the buffer from your median overrun.

## Frequently asked questions

### How long should a deep block be?

There is no verified universal number, so treat any figure, including ours, as a starting point. Over the first week, try 45, 60 and 90 minutes on similar work and note the minute you first reach for another tab or re-read the same line. Keep the longest length you finish before that happens.

### Does time blocking actually improve output?

We did not find a controlled study comparing time blocking with a to-do list. Newport has [written that](https://calnewport.com/the-time-blocking-revolution-begins/), in his experience, time blockers accomplish roughly twice as much work per week as people who use more reactive methods; he cites no study. The closest research is a [2021 meta-analysis of 158 studies](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0245066) of time management in general. It found time management moderately related to job performance (correlations around .25) and to wellbeing; its authors caution that most studies are cross-sectional, so causation cannot be established in all of them.

### Should I block the whole day or only the deep work?

Start with fixed meetings, the first deep block, one shallow batch and one buffer. Newport plans every minute of his workday, but he also suggests small deeper-work blocks at the edges of the schedule on mostly reactive days. Add blocks only when your log shows open time vanishing without a result.

## Sources and further reading

- [Exploring the "Planning Fallacy": Why People Underestimate Their Task Completion Times](https://doi.org/10.1037/0022-3514.67.3.366) — Buehler, Griffin and Ross, *Journal of Personality and Social Psychology*, 1994. Studies 1 and 4.
- [Maker's Schedule, Manager's Schedule](https://paulgraham.com/makersschedule.html) — Paul Graham, 2009.
- [Deep Habits: The Importance of Planning Every Minute of Your Work Day](https://calnewport.com/deep-habits-the-importance-of-planning-every-minute-of-your-work-day/) — Cal Newport, 2013. Evening planning, room for corrections.
- [Deep Habits: Three Recent Daily Plans](https://calnewport.com/deep-habits-three-recent-daily-plans/) — Cal Newport, 2015. Rewrites through the day.
- [The Time Blocking Revolution Begins...](https://calnewport.com/the-time-blocking-revolution-begins/) — Cal Newport, 2020. The 2013 reference and productivity claim.
- [Does time management work? A meta-analysis](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0245066) — Aeon, Faber and Panaccio, *PLOS ONE*, 2021.
- [Consider it done! Plan making can eliminate the cognitive effects of unfulfilled goals](https://doi.org/10.1037/a0024192) — Masicampo and Baumeister, *Journal of Personality and Social Psychology*, 2011. Plans and intrusive thoughts.
- [Use focus time in Google Calendar](https://support.google.com/calendar/answer/11190973?hl=en&co=GENIE.Platform%3DDesktop) — Google Calendar Help. Steps, defaults, plan requirements.