There is a genre of advice that begins with tracking your time. You install something, it charts your week into coloured blocks, and at the end you learn that you spent four hours in your inbox. You already knew that. Nothing changes.
The tracking is not the problem. The measurement is. Duration tells you where your hours went; it does not tell you whether they should have gone there. Those are different questions, and only the second one leads anywhere.
So this audit adds one column. It is a small change and it produces an uncomfortable number.
The Attention Ledger
For seven days, log every claim on your attention in three columns.
| What claimed it | How long | Chosen in advance? |
|---|---|---|
| Slack, design thread | 12 min | No |
| Drafting the proposal | 40 min | Yes |
| Checked phone, no reason | 4 min | No |
| Colleague asked about invoicing | 8 min | No |
The first two columns are an ordinary time log. The third is the method.
Chosen in advance means: if someone had shown you your calendar that morning and asked whether this thirty minutes should go here, would you have said yes? Not was it necessary. Not was it my fault. Only whether you would have chosen it, knowing what you knew that morning.
Most people find the answer is no far more often than they expect, and that the noes cluster in ways they did not expect either.
Four rules that make it work
Log at the moment of the switch, not at the end of the day. A reconstructed day is a story about a day. It smooths out the small switches, which is exactly where the finding lives.
Log your own switches too. The ones nobody caused. These are usually the majority and almost always the surprise.
Do not tidy anything. No rounding to the nearest fifteen minutes, no merging of three interruptions into one entry. The fragmentation is the data.
Do not change your behaviour during the week. You are taking a reading, not running an intervention. Improving your week while measuring it produces a measurement of a week you do not normally have.
The number
At the end of seven days, total the minutes marked no.
That figure is your unchosen hours: the portion of your working week that was spent without a decision. Not wasted, necessarily. Some of it was genuinely required. But none of it was chosen, and that is a different problem from waste and needs a different response.
Write it down as a fraction of your working week. Most people we have described this method to report being surprised by the size of it, though we want to be precise about what that is worth: those are reports from conversations, not a study, and we have not measured it. That distinction matters and this article will keep making it.
Reading your ledger
The total is the shock. The classification is the useful part.
Go through every unchosen block and mark it as one of three kinds.
| Class | What it means | The fix |
|---|---|---|
| Refusable | You could have said no, and the cost of saying no was survivable | A boundary, or a sentence you rehearse |
| Batchable | It genuinely needed doing, but not then | A window, later, where these go |
| Preventable | It should never have reached you at all | A setting, a default, a structural change |
These three need entirely different responses, and confusing them is the most common reason a focus attempt fails. People try to refuse things that were actually preventable, which means fighting the same battle every day instead of fixing it once. Or they try to prevent things that were genuinely refusable, which means building a wall against a person when a sentence would have done.
Start with the preventable ones
Refusable claims cost willpower every time. Batchable claims cost a decision every time. Preventable claims cost one afternoon, once, and then stop arriving.
A notification that should never have been on. A recurring meeting nobody defends. A channel you are in for historical reasons. An app that opens on a feed. Each of these is a permanent fix disguised as a daily annoyance.
Work the preventable list first, not because it is the biggest — usually it is not — but because it is the only class that compounds in your favour.
What the research suggests you will find
Two findings from the research literature are worth having in mind while you read your own ledger. Both are observed — they come from published work which we name below — and both describe populations rather than predicting what will happen to you.
Switching leaves something behind. Sophie Leroy’s 2009 studies named the effect attention residue: after moving from one task to another, part of your attention stays with the first, and performance on the second suffers. The effect was strongest when the first task was left unfinished or was under time pressure. This is why a two-minute interruption is not a two-minute cost — the recovery is not in the ledger, but it is in your afternoon.
Attention has been fragmenting for two decades, and we do it to ourselves. Gloria Mark’s group at UC Irvine has observed knowledge workers in real workplaces since the early 2000s, and reports that the average time spent on a screen before switching has fallen substantially over that period. The finding that complicates the story is that a large share of interruptions are self-initiated. Removing external interruptions does not empty the ledger; it reveals how much of it was you.
Where the evidence stops. These are findings about groups of people in specific workplaces over specific periods. They are not a diagnosis of your week, and no single study establishes what your unchosen hours ought to be. What the research does establish is that the phenomenon is real and measurable, which is the only claim this article needs it to support. Everything after this point is Eonara's interpretation.
The morning after
A ledger is a description. What follows is a decision, and it is smaller than people expect.
Do not redesign your week. Take the single largest preventable claim and remove it — the setting, the subscription, the default. Then take the largest batchable class and give it one window a day, at a time when your thinking is worst anyway.
Then protect one hour.
One. Not a morning, not a system, not a new calendar. The reason is not modesty. It is that an hour is the smallest unit you can defend without negotiating with anyone else, and the only way to find out whether your problem is the day’s structure or something else is to run one clean hour and see what happens in it.
If the hour works, you have found your unit and can build from there. If the hour does not work — if you sit down and cannot start, or start and cannot stay — then your problem is not the ledger, and the ledger has told you something almost as useful: the failure is somewhere else.
The audit ends. The deciding does not.
Seven days is enough. Do not keep the ledger running; a permanent audit becomes another claim on your attention, which is a joke the universe enjoys more than you will.
What survives the week is the question, asked in advance instead of afterwards: would I have chosen this?
That question is the whole of the first position on the Arc — the moment something asks for your time, before any decision has been made about it. Nearly everything that goes wrong with attention goes wrong here, because most claims are never examined at all. They simply arrive and are honoured.
Attention is the form time takes when you spend it. The hours arrive whatever you do. The only thing genuinely available to you is what they are spent on — and, having decided, whether the decision survives the day.
The next piece in this series is about that second half: why the ability to decide and the ability to stay are separate skills, and why almost all focus advice trains the wrong one.