Prioritizing Tasks: Why Urgency Wins Even When You Know Better

It is 4:47 on a Thursday. Two things are open on your screen. One is the piece of work that actually decides how your quarter goes. The other is a form that has to be back by five, and which nobody will ever think about again.

You know which one matters. You do the form.

Almost every article about prioritising tasks is written as though that were a knowledge problem, as though you did the form because nobody had yet shown you a grid with four boxes in it. You have seen the grid. You could draw the grid. You did the form anyway.

Which means the interesting question is not how to sort a list. It is why sorting it doesn’t stick.

The experiment that explains your afternoon

In 2018, three researchers ran five experiments to find out whether urgency has a pull of its own, separate from anything urgency is normally a proxy for.

That separation is the hard part, and it is why the study exists. Normally there are perfectly good reasons to do the urgent thing first: it is smaller, it is closer to done, the payoff is sooner and more certain, and you can always get to the important thing afterwards.

So the design stripped those reasons out. Participants chose between tasks where the objectively better payoff was known and on the table, and the competing task was worth less, but carried a deadline that was about to expire.

People took the worse deal. Repeatedly, and knowingly, in a way the authors describe as violating “the basic normative principle of dominance — choosing objectively worse options over objectively better options.” Their summary of what is going on is worth reading slowly: “People behave as if pursuing an urgent task has its own appeal, independent of its objective consequence.”

Not a preference for easy work. Not a failure to know what matters. A pull exerted by the expiring thing, on people who had the better option in front of them in numbers. That is why knowing your priorities has never been enough on its own: the ranking and the pull are two different systems, and only one of them is reading your list.

What happens at your house?

Which is why a better matrix will not save you

The most searched prioritisation tool in the world is the Eisenhower matrix: urgent and important on two axes, four boxes, do / schedule / delegate / delete. It is searched about fifty thousand times a month, and it has an origin story that is worth actually checking.

Eisenhower did say the line. On 19 August 1954, addressing the Second Assembly of the World Council of Churches in Evanston, Illinois, he said this:

This President said, “I have two kinds of problems, the urgent and the important. The urgent are not important, and the important are never urgent.”

Dwight D. Eisenhower, Second Assembly of the World Council of Churches, Evanston, Illinois, 19 August 1954

Read it again. He is quoting somebody else, “a former college president,” he says a sentence earlier, and he is not describing how he runs his desk. He is making a point about the Cold War, and what he says next is that this “represents a dilemma of modern man.” The matrix was drawn by other people, much later, out of one borrowed sentence in a speech about something else.

That is a piece of trivia. This is not: nobody has run a controlled trial on the matrix. Nor on ABCDE, MoSCoW, the Ivy Lee method, or any of the other four-box, five-letter systems that fill the first page of results. They are not treatments that have failed to work. They are organising devices that have never been tested, and they all do the same job, which is to hand you back a sorted version of what you already knew at 4:47.

The sorting was never the broken step.

What has actually been tested

There is one intervention in this area with a serious evidence base behind it, and it is almost never what productivity articles are about, because it is unglamorous and takes eleven seconds.

It is called an implementation intention, and it is a sentence in a fixed grammar: if it is [this time] and [this place], then I will [this specific action]. Not a goal. A trigger and a response, decided in advance.

d = .65

The pooled effect of if-then planning on rates of goal attainment, across the 94 studies in Gollwitzer and Sheeran’s review — a medium-to-large effect. Broken down: d = .61 on getting started, d = .77 on not getting derailed once you had.

The oddest finding in that literature is how much of the effect sits in the grammar. In one study, every participant was given the same four tasks, the same instruction to do them on Wednesday mornings, and the same freedom to pick their own time. The only difference was the sentence they filled in.

One group completed “I will perform as many arithmetic tasks as possible each Wednesday at ___.” The other completed “If it is Wednesday at ___, then I will perform as many arithmetic tasks as possible.” Their computers recorded when they actually started, and the conditional wording changed how closely people hit the time they had chosen.

Same intention. Same self-chosen slot. Different sentence, different behaviour. The reason appears to be that a conditional plan hands the decision to a cue rather than to you, so that when Wednesday morning arrives it is already carrying an instruction and there is nothing left to talk yourself out of.

There is a condition attached, and it is the one people break. The plan has to be precise enough that nothing is left to decide in the moment. Gollwitzer and Sheeran are blunt about the failure case: a plan that says “exercise more” in the then and “tomorrow” in the if has specified nothing, and leaves the person no better off than having simply intended to exercise more tomorrow.

“Get to the strategy doc at some point this week” is that plan. “If it is 9am on Tuesday and I am at my desk, then I open the strategy doc before email” is not.

You will also be wrong about how long it takes

You said the deck would take an afternoon. It is Sunday, the kids are watching something in the next room, and you are on your third pass at slide nine.

That is the second reliable failure, and it is arithmetic. People underestimate how long their own tasks will take, consistently, even when they have done the same task before and even when they know about the tendency. The name for it is the planning fallacy, and it has been studied for thirty years. It survives all that knowing because the estimate gets made from the plan, which contains no interruptions, rather than from the last four times, which were nothing but.

One experiment put both problems in the same room. A hundred and twenty students were asked to write a report within a set week. Half first wrote down where and when they would do it. The other half were simply given the task. Everyone then predicted how quickly they would finish.

The planners predicted 0.85 days and took 1.02. The others predicted 1.67 days and took 2.90.

Two things happened there, and only one of them is the obvious one. The planners were faster — nearly three times faster, in fact. But they were also right: their gap between prediction and reality, 0.17 days, was not statistically distinguishable from zero, while the unplanned group was out by more than a day. Deciding when and where did not just move the work. It made the estimate honest.

That is the practical version of the finding: your estimates are not a forecasting problem to be solved with better forecasting. They are downstream of whether the work has a time and a place.

The reason to bother is not the one you were sold

Here is where the honest answer diverges sharply from the genre.

The largest synthesis of this field pooled 158 studies covering 53,957 people, and asked what time management is actually associated with. Job performance: a correlation around .25. Real, modest, and roughly what a well-behaved psychological effect looks like.

Wellbeing did better. The relationship with psychological distress came out at −.358, and the authors’ own summary is that “time management seems to enhance wellbeing … to a greater extent than it does performance.”

That is not the promise on the cover of any productivity book. It is a better promise. The reason to get on top of your work is not that you will produce dramatically more of it. It is that the low-grade hum of having lost track of what you owe people is genuinely unpleasant, and it goes away.

One more thing from that review is worth knowing, because it quietly reframes this entire category. The field measures time management in three parts: structuring time, protecting time, and adapting when things change. Prioritising is not one of them. The part that gets almost no coverage, protecting the time once it exists, is the part carrying most of the weight.

A method that survives the evidence

What follows is not a system. It is the short list of moves that have something behind them, in the order they fail without.

Pick one thing. Do not rank the rest. You already know which one it is; you knew at 4:47. A full ranking of twenty items is work that produces a document, not a decision, and by tomorrow the list has changed anyway.
Give it a time and a place, in the if-then grammar. Not “Tuesday”. “If it is 9am on Tuesday and I am at my desk, then I open X before anything else.” If you cannot name the hour and the room, you have not made a plan, and the evidence says the vaguer version performs like no plan at all.
Decide now what you will say when someone asks for that hour. This is the protecting step, and it is the one the frameworks skip. Write the actual sentence: “I’m on the thing until eleven, can I come to you straight after?” Deciding it in advance is the same trick as step two, aimed at the interruption instead of the task.
Give urgency somewhere to go. The pull is real and it does not respond to being told it is unimportant. Put the small expiring things in one fixed block, late, after the protected hour, so they have a slot to be urgent in and stop competing for the one that matters.
Count what happened, not what you planned. Once a week, compare the two. This is the only cure for the estimating problem, and it takes five minutes.

Notice what is absent. No matrix, no quadrants, no letters, no ranking of the twenty. Every step is about when and where, because that is where the evidence is.

When everything on the list really is urgent

Sometimes the honest answer is that there is nothing wrong with your prioritising, and this is the part the genre almost never says out loud.

If four things are genuinely due, genuinely yours, and genuinely will not fit in the hours available, then you do not have a sorting problem. You have more work than time, and no arrangement of it into boxes creates an hour. Every method above will make that week marginally less awful and none of them will fix it, because it is not the kind of thing a method fixes.

What changes it is a conversation, and it goes better with a shape. Bring the list, not the feeling. Say what you can finish by when. Then ask the question that transfers the decision: which of these should slip? That is a different request from “I’m overwhelmed”, because it is answerable, and because it puts the trade-off in front of the person who is entitled to make it.

It also has to be said that this advice is not equally available to everyone. Some jobs punish the question, and knowing that the problem is structural is not the same as being able to say so. Where that is true, the honest version of this article is that steps two and three will protect one hour of your week, that this is worth having, and that it is not a solution to a workload.

The bottom line

The urgent thing is not winning because you are disorganised. It is winning because expiring things exert a pull that survives knowing better, which is exactly why the fix cannot be more knowing better.

So stop ranking and start scheduling. One thing, one hour, one room, one sentence ready for whoever comes to the desk. Then look at what actually happened on Friday, and let that, rather than your optimism, set next week’s number.

And if the payoff turns out to be that you sleep better rather than that you ship more — that is not the consolation prize. On the evidence, that was always the larger effect.

Sources (5)

Every factual claim in this article comes from one of these sources.