You give a task “the morning” and it takes the morning, then the afternoon, and it still isn’t finished. The task had no ceiling, so it kept absorbing polish, research and second-guessing.
Timeboxing makes time the constraint and scope the variable. Before you start, you set the maximum time a task gets and decide what “done enough” looks like. When the box closes, whatever meets that bar ships.
This guide covers how timeboxing differs from time blocking, when to use hard or soft boxes, how to size them, and what to do when the timer beats you.
Where timeboxing comes from and why the cap works
Timeboxing started in software project management. The DSDM method, now stewarded by the Agile Business Consortium, formalized it as fixing the time for a piece of work and letting the delivered features flex. The deadline never moves; the amount done inside it does.
Scrum works the same way. The Scrum Guide defines its events as timeboxes with a maximum duration, such as 15 minutes for the Daily Scrum. That number is a ceiling, not a target.
Marc Zao-Sanders, author of Timeboxing: The Power of Doing One Thing at a Time, brought the idea to personal calendars, where each timebox is a calendar entry with one task and a defined end.
The psychology behind the ceiling is simple. Open-ended work invites what Barry Schwartz and colleagues called maximizing, the hunt for the best possible result. Their research linked maximizing to more regret and lower satisfaction than “satisficing”, which means accepting an option that meets a clear standard.
A hard end time forces you to write that standard down before you start. It also blocks the drift described by Parkinson’s Law, where work expands to fill whatever time you give it.
Timeboxing is not time blocking (or Pomodoro)
Cal Newport’s time blocking gives every minute of the workday a job and lets you redraw the plan as the day changes. It answers “when will I work on this?” Timeboxing answers a different question: “how much is this task allowed to cost, and what will exist when that time is up?”
The two combine well. A time-blocked calendar might hold a 9:00 to 11:00 slot for “proposal work.” Timeboxing turns that slot into “9:00 to 9:45: draft pricing section, good enough to send to finance.”
If you already follow our guide to time blocking, think of timeboxing as the discipline you add inside the blocks.
| Timeboxing | Time blocking | Pomodoro | |
|---|---|---|---|
| Core mechanism | A fixed maximum time plus a defined output. Scope flexes, time doesn’t. | Calendar slots for every part of the day, redrawn when plans change. | Repeating 25-minute work intervals with short breaks (Francesco Cirillo). |
| When the timer rings | Ship what meets the bar or re-scope (hard). Make one pre-agreed decision (soft). | Move to the next block, or revise the blocks that are left. | Take a 5-minute break, then often continue the same task. |
| Best task types | Drafts, decisions, research, admin batches, anything prone to perfectionism. | Days that mix meetings and several projects needing protected time. | Tasks you’re avoiding; long stretches of steady, repetitive work. |
| Common failure mode | Boxes named after an activity, with no done bar. | The whole plan is abandoned after the first disruption. | The interval cuts into deep work just as it gets going. |
Pomodoros can run inside a timebox. A 45-minute drafting box can hold a pomodoro and a half if short intervals help you start. The box still owns the deadline and the definition of done.
Hard timeboxes vs soft timeboxes
A hard timebox ends when the timer rings, with no exceptions. You save, ship or hand over whatever exists.
Use hard boxes when the output has a natural “good enough” bar, when perfectionism is the main risk, or when other work depends on you stopping on time. Status updates, first drafts, inbox processing, research spikes and decisions all fit.
A soft timebox ends with a decision instead of a stop. When the timer rings you choose one of three moves: stop, extend once by a fixed amount (15 minutes is a sensible default), or re-scope and book a new box.
You get one extension only. Needing a second one means your estimate was wrong, and that is information to record, not a reason to keep going.
Soft boxes suit work where stopping mid-thought is costly, such as debugging, a tangled analysis or the final pass on a client deliverable. Mark each box H or S in the calendar title. That way the rule is set before you start, not argued over at the moment you most want to keep going.
Sizing your boxes: 15, 45 and 90 minutes
Three standard sizes cover most knowledge work. Fewer options make planning faster and give you cleaner data on how long things really take.
| Size | Use it for | Define done as | Default type |
|---|---|---|---|
| 15 min | Admin: invoices, scheduling, approvals, a batch of short replies | A count: “clear 8 flagged emails” | Hard |
| 45 min | Drafting: a section, a slide outline, a spec, a first pass at code | A rough but complete artifact | Hard |
| 90 min | Deep work: complex writing, analysis, design, architecture | A milestone: “model runs end to end” | Soft |
Cap boxes at 90 minutes. A task that needs more than that usually contains two outputs, so split it and box each one.
Add a 10-minute gap after a 90-minute box. Our guide to managing breaks covers what to do with that gap.
Here is how the sizes play out. A hospital finance controller has the monthly board finance report due Thursday. Instead of writing “Wednesday: work on report,” they book a 15-minute hard box to request missing activity figures from surgery and outpatients.
Then come two 45-minute hard boxes to draft the patient-income and staffing-cost commentary, and one 90-minute soft box to build and check the variance charts. That is 3 hours 15 minutes, each piece with a visible finish line.
Write the actual time next to each box afterward. After about ten boxes of one size, compare actual time to planned time. If your 45-minute drafting boxes usually need 60, either plan 60 or make the output smaller.
When the timer rings and the task isn’t done
This is the moment that decides whether timeboxing works for you or falls apart. Run this sequence in the two minutes after a box expires with work left over.
- Check against the done bar. If the output meets the bar you wrote in the title, it ships, even if you’d like another pass.
- Re-scope with MoSCoW. The MoSCoW method sorts remaining work into Must, Should, Could and Won’t have. When time is fixed, the Coulds go first. List what’s left, label each item, and carry only the Musts forward.
- Book the follow-up box now. Make it smaller and give it a new output, such as “45 min: finish pricing table, Musts only.” Avoid vague titles like “continue proposal.”
- Write a two-line re-entry note. Note where you stopped and the very next action. This closes the loop so the unfinished task is less likely to follow you into the next box.
For example, a UX writer books 45 minutes for onboarding copy across six screens and finishes four. Screens one to four meet the bar and go to design review.
Of the two left, the payment screen is a Must and a tooltip is a Could. The next box is 15 minutes for the payment screen only, and the tooltip goes to the backlog.
Why timeboxes leak
- Activity names instead of outputs. “Work on deck” has no finish line. Use verb plus object plus bar: “Outline deck: 10 slide headlines.”
- Silent extensions. Hard boxes quietly turn soft. Log every overrun. If more than a third of your hard boxes run over, resize them; the estimates need correcting.
- Wall-to-wall boxes. One overrun wrecks the rest of the day. Leave about a fifth of the day unboxed to absorb overruns and requests.
- Interruptions that dissolve the box. Defer with a specific time: “I can look at that at 11:15.” Our guide to saying no has more wording for this.
- Boxing everything. Save boxes for work that sprawls. Routine tasks with obvious endpoints don’t need one.
Six fields for every box
Fill in these six fields for each box, either in the calendar event or on a card next to your keyboard.
- Title: verb + object + done bar, for example “Draft Q3 risk section: 600 words, sources linked”
- Length: 15, 45 or 90 minutes
- Type: H (stop at the bell) or S (one 15-minute extension)
- Must-haves: the two or three things that make the output shippable
- If it expires: who gets the partial output, and when the follow-up box happens
- Actual time: filled in afterward to sharpen your estimates
Start with one hard 45-minute box tomorrow on the task you’ve been polishing longest, and ship what exists when it ends.
6 Comments