When everything is high priority: why the label stops deciding for you

Why is everything on my to-do list marked high priority?

Open the app and sort by priority. Thirty-one tasks under High, four under Medium, and a Low tier holding two things you are never going to do and know it. The sort ran, the list reordered itself slightly, and what you are looking at is almost exactly the list you had before. Nothing got decided.

That is a consequence of when you set the label rather than of how carefully you set it. A priority goes onto a task at the moment you capture it, and at that moment the task is alone on the screen. The only question available is whether this thing matters, and for work you bothered to write down the honest answer is usually yes. High is what that yes looks like in a dropdown.

Ranking needs a comparison and labelling does not. Tagging one task High is not a claim that it beats the other thirty, it is a claim that it is not trivial. Do that fifty times across three weeks and the top tier fills with work that is, one item at a time, not trivial. Which is the right answer to a question you were not trying to ask.

Then nobody ever goes back and marks anything down.

Downgrading has no trigger. A task gets graded once, at capture, and keeps that grade until it is done or deleted, because no moment in your week asks you to look again. The urgent thing from a fortnight ago still reads High although the person who asked answered their own question on Thursday. So the tier is not only crowded, it is stale, and stale entries are the ones you learn to skim past, which teaches you to distrust the field you were trying to use.

The cost lands on the days you most need a ranking. A quiet Tuesday does not need a priority field at all, since you can hold the shape of that day in your head. A Tuesday with two escalations and a deadline is when you open the list wanting it to tell you something, and that is precisely when thirty-one High tasks tell you nothing.

Do priority levels in a to-do app actually work?

Mostly they have been studied somewhere else. No research exists on the High, Medium and Low field in your task app, and a page that suggested otherwise would be inventing it. What does exist is a body of multitasking work asking a narrower version of your question: when a person is told that one task matters more, does that change which task they actually pick? The results are mixed, but the answer came back no often enough that the finding earned a name. Researchers call it priority neglect.

That finding picked up a condition in 2024. Lewis, Gutzwiller and Johnson, writing in Applied Ergonomics, put operators through the Multi-Attribute Task Battery, a simulator that throws concurrent tasks at you the way a monitoring console does and considerably faster than your Tuesday does. They ran it twice, once with an auditory communications task as the priority and once with a manual tracking task. Instructed priority did significantly increase selection of the target task, and it did so under the conflicting conditions: the moments when doing the priority task meant not doing the other one.

Read that back onto your own list and the shape is familiar. A label earns its keep at the point where two pieces of work want the same hour. It does close to nothing for a task sitting by itself in a backlog, which is where almost all of your labels were applied. The abstract reports no effect size, and the study watched trained operators at a console rather than anyone with your list open, so what carries across is the condition to look for, not a number to quote.

A second study says something about which of your days a label survives. Valery and colleagues, publishing in the same journal in 2019, gave twenty participants two N-back memory tasks to run at once, each carrying its own difficulty level and its own point value. When the combined difficulty was low, people spent their attention on whichever subtask was harder. When it was high, they spent it on whichever subtask was worth more points, and the switch between those two strategies was not a gradual trade.

So a stated importance starts to bite as the day gets heavier, which is the reverse of what you probably assume about yourself under pressure. Hold it loosely. Twenty people, an N-back task, one lab.

The third piece is about you rather than about the task. Jansen, van Egmond and de Ridder ran three experiments pairing a high-speed driving game with an auditory memory task, published in PLoS ONE in 2016, and found that people turn up already preferring one task over the other. Priority instructions could overrule those preferences, but only after participants had spent more time in the dual-task situation. You have preferences of that kind too, formed by whichever sort of work tends to get noticed, and a label has to beat them before it changes anything.

None of that was measured on a backlog, and none of it hands you a rule for Monday morning. Taken together it points one way: a priority marking acts on your choice in the narrow case where two things genuinely conflict and you have been in that situation long enough to recognise it. Your list mostly does not meet that description, and that is a fixable property of the list.

How many tasks should be high priority at once?

Three, or whatever number you can honestly finish in a day, and the useful part is that a number exists at all. A tier with a cap is a ranking. A tier without one is a filing category that happens to be called High.

Set the cap and the label starts working immediately, because adding a fourth task now means taking one out. That is the comparison the capture moment never offered you. It costs a few seconds, and it is the entire mechanism: nothing gets promoted without something being demoted, so every promotion becomes a decision about a pair rather than a verdict on one task.

Pick a fixed moment to do it. The end of the day works well, since the work is behind you and no fresh request can land before tomorrow, which is the logic the Ivy Lee Method already runs on with six tasks written the night before. A Monday slot works if your week is planned in blocks rather than in days. What does not work is re-ranking when you notice, because the moment you notice is the crowded Tuesday afternoon, and that is when the decision is hardest and least honest.

Cap the top tier and the middle one quietly stops mattering.

Which is fine. Medium and Low were never doing much beyond recording that you had once looked at a task. If your list holds forty things and three of them are today, the other thirty-seven need a sort order and a date to look again, not a grade. Plenty of apps offer that as a manual order, and an order beats a grade, because an order contains no ties and a three-tier grade is made of almost nothing but ties.

The case a cap cannot solve is two tasks that both have to be first, with no exaggeration in either claim. That is the real conflict those multitasking studies were staging, and no label breaks it, because the tie sits in the work rather than in the app. Something moves. The question worth asking is which outcome you would rather explain.

A method helps here where a label cannot. The Eisenhower Matrix suits a tie between a task with a deadline and a task that matters, since it forces those two questions apart instead of collapsing them into one field. When the same tie turns up every week, the problem is how much work you have open at once rather than how it is graded, and the work-in-progress limit that Personal Kanban is built around will do more than any relabelling. And when the whole list carries deadlines other people set, the audit on the prioritization page is the piece you actually need.

How does your working style change the answer?

Where the top tier goes wrong depends on how you already work. The four names below are this site's own, from the working-style self-check on the home page.

Architects tend to have the cleanest labels and the least appetite for moving them. A task was graded when it was planned, the plan was sound, and re-grading it feels like conceding that the plan was wrong, so an Architect's High tier goes stale rather than crowded. Sprinters inflate theirs fastest. Anything worth a burst gets marked High during the burst, and the burst ends well before the label is revisited.

Visionaries fail differently again. The work that matters most to them usually carries no deadline and no requester, so it sits at High for months without moving, alongside eleven things that are High because somebody asked this week. Improvisers may be who the cap was built for. They label least and reorder most, which works while the week stays calm and leaves them nothing to fall back on when three things land at once.

Some lists this page does not describe. If your priority field is set by a project manager or a ticketing system rather than by you, the tier is a communication channel, its inflation is somebody else's to fix, and what you control is the order you work within a tier. If the top tier is honestly full because four people are each waiting on something today, no cap invents capacity for you. That one is a conversation about what moves, not a labelling problem, and the earlier it happens the more of it is still your choice.

So open the list, sort by priority, and count what is sitting at the top. Pick the three you would defend if somebody made you justify the list out loud, knock everything else down a tier, and leave it there until tomorrow evening. That count is the measurement worth keeping. If the top tier is back over ten by Friday, the cap is not what failed; the intake is.

Frequently asked questions

Sources

• Lewis, Gutzwiller & Johnson (2024), Applied Ergonomics: Priority influences task selection decisions in multi-task management
• Valery et al. (2019), Applied Ergonomics: Global difficulty modulates the prioritization strategy in multitasking situations
• Jansen, van Egmond & de Ridder (2016), PLoS ONE: Task Prioritization in Dual-Tasking: Instructions versus Preferences