TL;DR. Estimate your 24 hours from memory first, in five minutes, with nothing open in front of you. Then check it against one honestly observed day. The gap between the two pictures is the entire lesson, and it usually shows up in the same three places: sleep counted generously, transitions counted at zero, and phone time missing. One day is a weak instrument and I will say so twice, but it is enough to break a false picture and get you to one change. It is also the version you will actually finish.
Ask me to describe my week and I will give you a clean answer. Contract work through the day, an hour or two on the app in the evening, gym when I can, time with my daughter, sleep around seven hours.
That description is a lie assembled entirely out of true statements. Everything in it happened. What is missing is everything that has no name: the twenty minutes it takes me to actually start after the laptop is open, the phone check that runs long while the kettle boils, the fifteen minutes of standing in the kitchen after bedtime doing nothing in particular. Those are not in my mental picture of the day, because they were never labelled well enough to be stored.
The exercise below closes that gap in about five minutes, and it is deliberately the light version. If you want the thorough one, run a proper multi-day time audit instead, which is three to seven days logged in 30-minute blocks. Most people will not do that. This is for them.
Step one: draw the day from memory
Do this part before you look anything up. That instruction is the whole design.
Open the daily time pie and assign your 24 hours across the categories: sleep, work or study, commute, screens and TV, chores and errands, meals and cooking, family and social, exercise, self-care, free time. Add your own if the presets miss something, up to eight of them. Hours move in quarter-hour steps and a counter shows how much of the day you have left to allocate, so the budget is enforced: you cannot describe a 27-hour day, which is a constraint most people meet for the first time here.
Do not check your phone. Do not open your calendar. Do not adjust a number because it looks bad. You are recording the day as you believe it to be, and a corrected guess measures nothing.
When it sums to 24, take the text summary and copy it somewhere. You will need it in step three, and the tool stores one actual day at a time, so the next version overwrites this one.
Step two: check one ordinary day
Now pick a day. Not a special day, not next Monday when everything will be better organised. Today, or tomorrow.
You are not logging every activity. Three checkpoints are enough:
- After lunch. Write down roughly what the morning contained.
- At the end of work. Same for the afternoon.
- Before bed. The evening, plus one number you do not have to remember: your phone's own screen time report.
That last number is the only hard data in the whole exercise and it is free. Every phone reports it. Take the figure it gives you rather than the figure that feels right.
Then rebuild the pie with what the day actually contained.
What people get wrong in the guess
Three errors come up so often that you should brace for them. All three are structural rather than moral, which is why the same people make them every time.
Sleep gets over-reported. Most of us report time in bed. The gap between lights out and asleep, the twenty minutes on the phone first, the stretch awake at 3am, all get counted as sleep because they happened in bed. If you told the guess eight hours, the honest number is frequently closer to seven.
Transitions come out at zero. Nobody has ever estimated their day and included the time it takes to start something. Getting ready, walking to the car, the reorientation after an interruption, the drift between finishing one task and beginning the next. Each one is too small to remember and together they are often the largest unassigned wedge in the picture.
Phone time is missing. Not underestimated. Missing. Scrolling in a queue, in the kitchen, in the two minutes before a meeting starts does not feel like an activity you began, so it never gets stored as one. This is why the screen time report is worth more than your memory, and why the honest pie usually has a much bigger red slice than the guessed one. If that number lands badly, how to reduce screen time is the practical follow-on, and the screen time calculator will run your daily number out over a year and a lifetime if you want the uncomfortable version. Fair warning about that one: it multiplies your inputs forward and its comparison cards use fixed assumptions, such as counting a book as six hours. It is arithmetic on your numbers, not a forecast.
How to read the picture without spiralling
This is where the exercise goes wrong for people, so it is worth being explicit.
Sleep is not a wasted slice. It is the largest single input into how every other slice goes. A pie with eight hours of sleep is not a pie with eight hours of missed opportunity.
Care work is not a wasted slice either. Time spent with a child, a parent or a partner is not overhead in a productivity chart. My own pie has a large wedge that no optimisation is ever going to touch, and I would not want it to. If the exercise makes you feel guilty about the hours you spend on people, you have read the chart backwards.
Chores and admin are load-bearing. They can often be batched and sometimes reduced. They cannot be zero, and treating them as failure just adds shame to laundry.
The chart has no moral column. It has a comparison against your own ideal day, which is the only comparison that means anything here. Build that second pie in the tool's ideal mode, put both side by side, and read the per-category gap. There is also a set of rows comparing your day against a US average derived from the American Time Use Survey, which is mildly interesting context and not a target. Nobody lives at the national mean.
The two outputs I find genuinely useful are the biggest-gap callout and the yearly projections, which turn a daily slice into full days and full weeks per year. A daily half hour is trivially dismissable. The same number described as a fortnight a year is not.
Then change one slice, once
Here is the part where most self-audits die: you see the picture, feel a burst of clarity, plan a redesign of your entire life, and by Wednesday you are exactly where you were with an added layer of guilt.
Change one thing.
- Take the biggest gap between actual and ideal. Not the most embarrassing slice. The biggest gap, which the tool names for you.
- Halve it, do not eliminate it. Halving survives a bad week. Elimination is a promise you break, and breaking it usually ends the whole project.
- Give the recovered time a destination the same day. Time freed with no destination is reabsorbed within about a week, almost always by the thing you just cut. If you do not know where it should go, how to plan your day is the ten-minute ritual that decides that in advance.
- Re-run the pie in a month. Same five minutes, from memory again first. The comparison against your old text summary is the actual finding, which is why step one told you to save it.
What a one-day snapshot cannot do
I said I would say this twice.
One day is a weak instrument. It cannot distinguish a typical Tuesday from an unusual one. It cannot see the weekend, which has a completely different shape and is often where the recoverable hours actually live. It flattens a week with three different kinds of day into a single average that describes none of them. A sick child, a deadline, or one late night distorts the whole picture, and you will not know which of those you caught.
So do not treat the output as a measurement of your life. Treat it as a test of your assumptions.
That is a narrower claim and still a useful one, because the picture you carry in your head is what you make decisions with, and if it is wrong by two hours in the wrong direction then every plan you build on it inherits the error. Five minutes is enough to find out. A snapshot that gets taken beats a rigorous protocol that does not.
If the snapshot leaves you wanting the real number, that is the moment to run the full multi-day version, and you will run it with a specific question rather than a vague sense of dread, which is a much better way to start.
And if what the pie makes you feel is not "where did today go" but "where did this decade go", the zoom-out version of the same exercise is your life in weeks. Same medicine, longer scale, harder to look at.
Five minutes, now
Guess your day. Save the summary. Check one ordinary day against it. Change one slice by half and give the time somewhere to land.
Everything stays in your browser and there is no account. The only difficult part is the part no tool can do for you, which is entering the real number instead of the flattering one.
Last updated: December 2026.
Written by Eusebiu, the solo founder building Loggd in public. I build the app in the evenings around contract work and a small daughter, and the first time I drew this pie honestly the gap between the evening I thought I had and the evening I actually had was close to an hour. I share what the aggregate data keeps showing on Threads.
Seeing the slice is not the same as changing it. Loggd tracks the habits, tasks and focus sessions you protect after an exercise like this on a forgiving contribution grid, so the change is still visible six weeks later. Start free.
