Sep 19, 2026 · 6 min read
Ask someone how long it takes them to get ready in the morning and they will say twenty minutes. Stand next to them with a stopwatch and it is closer to forty. The gap is not a character flaw; it is the normal way people estimate, and it is fixable in an afternoon with two tools: a stopwatch to find out, and a countdown to hold the line once you know.
Daniel Kahneman and Amos Tversky described the planning fallacy in 1979: people estimating a task lean on the best-case story of how it will go and leave out the delays, interruptions, and small extra steps that nearly always turn up. You estimate the writing, not the finding of the file, the reading back, the fixing of the one paragraph that would not sit right. The estimate is for the task; the real duration is for the task plus everything around it.
Two habits help. First, estimate from the outside: not "how long will this take?" but "how long did this take last time?" Your last three attempts are better evidence than your intentions. Second, estimate the pieces, not the whole. "Clean the kitchen" invites a single optimistic number; "empty the dishwasher, load it, wipe the surfaces, sweep" invites four honest ones that add up to more.
The order matters. A countdown set from a guess teaches you nothing except that the guess was wrong. So the first time through a task, do not time it at all. Measure it.
The stopwatch here is drawn as growth rings rather than a shrinking dial: the first ring takes ten seconds, and each ring after takes twice as long, so short measurements fill the centre quickly and a long one keeps growing. It is anchored to the clock, so a background tab or a laptop that went to sleep still reports the honest elapsed time when you come back.
Once you know what a task really takes, you have a choice. You can give it that long, or you can give it less and see what happens. Time-boxing is the second option: you set the duration before you start and stop when it ends, finished or not.
It works for two kinds of tasks. The first is the open-ended kind, where "done" is a judgement call: tidying, research, editing, planning a trip. These expand to fill whatever you give them, so give them a box. The second is the dreaded kind, where starting is the hard part. A box of fifteen minutes is a bargain you can accept, and most of the time the task keeps going after the chime.
A time-box is only honest if the end is real. That means a countdown you can see, an alarm you will hear, and a rule that when it rings you stop or consciously decide to set another. Every timer on radialtimers.com chimes at zero and keeps running in a background tab, so the box holds even while you work in another window.
These are typical ranges for an unhurried adult, meant as starting points for your own measurements rather than facts about you. The point of the table is the width of the ranges: nearly every everyday task varies by a factor of two depending on the day. Measure yours, then pick a timer from the upper half of your range.
| Task | Typical range | What stretches it |
|---|---|---|
| Morning routine, wake to door | 30–60 min | Breakfast, deciding what to wear, phone |
| Shower | 5–15 min | Washing hair, no timer |
| Brushing teeth | 2–3 min | Rarely stretches; usually cut short |
| Making a simple breakfast | 5–15 min | Anything that needs a pan |
| Cooking a weeknight dinner | 25–60 min | Chopping, more than one pan, a recipe you have not made before |
| Washing up after dinner | 10–25 min | Pans, no dishwasher |
| A load of laundry, start to folded | 2–3 h | Dryer cycles, forgetting the load |
| Cleaning a bathroom | 15–30 min | The shower screen |
| Vacuuming an apartment | 15–30 min | Moving furniture, stairs |
| Writing one considered email | 5–20 min | Reading the thread first, second thoughts |
| Clearing an inbox of 50 messages | 30–90 min | Any message that needs a reply |
| Reading 20 pages of a novel | 20–40 min | Dense prose, a phone within reach |
| Reading 20 pages of a textbook | 45–90 min | Notes, working the examples |
| Walking a mile | 15–22 min | Hills, crossings, company |
| A grocery shop, door to door | 45–90 min | Travel, a long list, the queue |
| Packing for a weekend away | 20–45 min | Not having a list |
| Preparing a ten minute talk | 2–4 h | Slides, rehearsing aloud |
| A haircut, walk in to walk out | 30–60 min | Waiting, a wash |
Notice how often the third column is a specific, predictable thing. That is the planning fallacy in a table: the stretch is never a mystery, it is just the part left out of the estimate.
A single measurement is a fact; a repeated one is a plan. Pick the five tasks you do most and measure each three times over a couple of weeks. Write the middle result next to each. Then, from a hub like every timer on one page, pick the standard duration just above it and use that timer every time. The task now has a slot, and slots get done.
Two effects follow. The first is that your estimates for unmeasured tasks improve, because you have a feel for how wide the ranges are. The second is that the timers become the plan: a morning is a 40 minute timer, a 15 minute timer for the walk, a 50 minute timer for the first block of work, and you stop negotiating with each one.
The site's stats page is public and shows the whole site rather than any one person: the hours everyone here has spent with a timer running, visits, timers started, and the number of people keeping time, with charts for the last thirty days, today against a typical day, and the pattern hour by hour. It is a good reminder that a lot of small timed blocks add up. Your own hours are kept on the leaderboard once you sign in, where they follow you across devices and count toward a streak.
The middle one, then round up. If three attempts at the same task gave 22, 31 and 26 minutes, plan on 30. Using the best case is how the planning fallacy starts again.
No. A countdown with a little slack, say ten percent, lets you finish without racing the last ring. A countdown with no slack turns every task into a sprint, and you will start ignoring the chime.
Find the closest task you have measured and double it. Then measure the new one with the stopwatch as you go, so next time you have a real number instead of a doubled guess.