The Eisenhower Matrix is good at one job, but it gets used for two. The job it does well is sorting: you take a crowded list, split the fires from everything else, and see what can wait until tomorrow. The job it does badly is deciding what deserves your hours this week and this quarter. People confuse the two because the grid looks as if it answers both questions. When you put a task in the box marked important, the grid seems to have told you the task matters. But the grid didn’t decide that. You did, when you chose the box.
So this guide keeps the matrix for the sorting and looks hard at the rest. It names five specific ways the matrix fails when you ask it to set priorities, and it gives one repair for each. After that it runs one manager’s Monday list through the plain version and the repaired one, so you can see what changes and what it costs. None of this means throwing the matrix out, though. The guide also covers where the critique stops being useful, because on some days the plain grid is exactly what you need.
What is the Eisenhower Matrix good for?
The matrix is sometimes called the Eisenhower Box. It sorts tasks on two questions: is it urgent, and is it important? Answering both gives you four boxes. Urgent and important work you do now. Important work that isn’t urgent you schedule for later. Urgent work that isn’t important you hand to someone else if you can. Work that is neither you drop. If you want a full walkthrough of the four boxes, our guide to the urgent and important grid covers them. Our step-by-step training plan for managers also uses the box as one of its steps.
What it’s good at is triage. When forty things have landed on you and they all feel equally loud, the grid gives you a cheap way to look at the whole pile at once. It takes a few minutes and needs no scoring or spreadsheet. It also usually shows you that a good part of the list can wait. That’s why it shows up in so many guides to prioritization. Our own roundup of prioritization frameworks is right to recommend it daily as a triage tool.
But this post is about the other job people hand it, which is deciding what matters enough to get your best hours. That’s a question about time management across a week or a quarter, and it’s where the grid starts to struggle.
Where does the Eisenhower Matrix actually come from?
The name points to Dwight Eisenhower, and there is a real Eisenhower quote behind it. On August 19, 1954, Eisenhower spoke to the Second Assembly of the World Council of Churches in Evanston. He told the audience he would illustrate a point “by quoting the statement of a former college president”. The transcript of Eisenhower’s 1954 Evanston address is kept by the American Presidency Project at UC Santa Barbara. After that line it reads: “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.’”
So the line wasn’t his, and he said so. Quote Investigator’s history of the saying reaches the same conclusion: “Dwight Eisenhower did present a version of the adage in a 1954 speech, but he credited an unnamed ‘former college president’.” The versions you usually see say the important is “seldom” urgent, but that isn’t what the transcript says. Quote Investigator offers two explanations. The softer wording may have come from a faulty transmission of the speech, or Eisenhower may have used more than one version of the line.
The 1954 wording says something odd if you read it literally. If the urgent are not important and the important are never urgent, nothing can be both. That would mean the urgent-and-important box is always empty. This is our reading of the words rather than a finding from any source. Even so, it shows the speech describes two kinds of problem, not a grid. No grid appears anywhere in his remarks either.
The four-box grid that people now call the Eisenhower Matrix appears in Stephen Covey’s 1989 book The 7 Habits of Highly Effective People. It’s in the chapter on Habit 3, where it’s called the “Time Management Matrix”. We won’t claim more than that about who drew it first. But one detail from Covey’s chapter is worth holding on to. He defines an important task by its results. His test is whether it moves you toward your mission, your values or your most important goals. That definition comes back in the second failure below.
What are the limitations of the Eisenhower Matrix? Five failures, one fix each
The five failures below are not equally well supported, so it’s worth knowing which is which before you weigh them. The first is that urgency is an unreliable input, and it rests on a published set of experiments. The other four are our editorial judgment, reasoned from how the grid is built and from Covey’s own description of it. The fixes are practical suggestions, not tested treatments. Each section says what goes wrong and why. It then says what to do about it and what the fix costs you.
Urgency is a feeling, not a measurement
The matrix asks you to rate how urgent each task is, as though urgency were a property you could read off the task. But often it isn’t. It’s a pressure the task puts on you instead. There’s good experimental evidence that this pressure can pull people toward a worse choice even when the better one is in plain view.
Meng Zhu, Yang Yang and Christopher Hsee published a set of experiments on what they called the mere urgency effect in the Journal of Consumer Research in 2018. They define task urgency as “the state that requires immediate responsiveness”. Their question was simple. If you offer people two tasks and one clearly pays more, does a short deadline on the other one pull people toward it? One experiment asked 203 online workers on Amazon’s Mechanical Turk to choose between two short typing tasks. One paid a bonus of 12 cents and the other 16 cents. When the 12-cent task came with a five-minute expiry instead of a fifty-minute one, the share of workers choosing it rose from 13.9% to 35.3%. Nothing about the task had improved. The only change was that it expired sooner.
A later experiment in the same paper ran three rounds with 402 workers. The prize was a chance in a lottery for a $20 or a $25 gift card. When the lower-value option carried an expiry, 58.8% of workers chose it at least once across the three rounds, against 38.4% when it didn’t. The authors also found the effect was “more pronounced among workers who perceived themselves as busy in general”. That matters here, because a sorting grid is a tool built for busy weeks.
It’s just as important to be clear about what the study doesn’t show. The stakes were cents, chocolates in one experiment and gift-card lottery entries in another. The tasks were short typing and review-writing exercises. The people were university students and online workers, not managers making decisions about their teams. Most people still picked the better task in each of the single choices the experiments report, even under a short deadline. The urgency was also invented for the experiment. It was an expiry that existed only to create pressure. A real deadline is different, and something like a payroll cut-off is a perfectly good reason to act first. So the study doesn’t tell you how often managers get priorities wrong. What it shows is that a short clock makes people more likely to pick an option they can see is worse.
That’s awkward for the matrix, because the matrix asks the same person whose attention is being pulled by the clock to rate the clock. The authors’ explanation is that urgency “elicits attention, diverting focus away from the task payoffs”. If that’s right, the urgency rating is bent before you’ve sorted anything.
Another experiment in the paper, with 402 workers, suggests a repair. When workers got no reminder of what each task paid, 33.7% took the lower-paying task if it was urgent, against 9.8% if it wasn’t. But in another condition the payoffs for both tasks were “highlighted at the moment of task choice”. With that reminder the gap shrank to 20.6% against 15.0%, and the difference wasn’t statistically significant. The reminder was only two short sentences restating payoffs the workers already knew.
So the repair is simple. Before you sort, write next to each task what actually happens if it waits. Say what goes wrong, for whom, and by when. That puts the cost of waiting in front of you at the moment you choose, which is the condition under which the effect shrank in the lab. But there’s a gap between that lab result and your desk. The workers only had to be reminded of a payoff they knew. At work you usually have to estimate what waiting costs, and estimates can be wrong. We haven’t found any test of this repair in a workplace, so treat it as a practical step built on the lab finding. It costs about a minute per task, and you can skip it for anything with an obvious external deadline.
”Important” has no yardstick
The matrix lets you call a task important without saying what it’s important for. “Important” drifts when nothing measures it. It ends up meaning whatever the most senior person asked for, or whatever you happen to be most anxious about at four in the afternoon. That’s how two people can sort the same list and both feel confident. Neither of them had to say what their labels were based on.
Covey’s own chapter points at this. The passage that introduces the grid also ties importance to results: a person’s mission, values and main goals. It warns that you’re easily pulled into whatever is urgent if you haven’t first done the work of Habit 2, which is getting clear on what matters and what results you want. The grid sits in the chapter on Habit 3, after that work is supposed to be done. But the version most people meet is the four boxes on their own, without the step that was meant to come first. That’s why a grid on its own can’t make sure important work gets attention. It can only hold the label you give it.
The fix is to write two or three outcomes for the next quarter before you sort anything. Make one or two for your team and one for yourself. Every task you mark important then has to name the outcome it serves. A task that serves none of them isn’t important this quarter, however loudly it arrived. This is goal setting doing a job the grid can’t do for you. The cost is that the first version takes real thought, and probably a conversation with your own manager about whether you’ve picked the right outcomes. After that you reuse the list every week until the quarter changes.
It ignores effort and capacity
A task in the “schedule” box might take twenty minutes or three weeks, but the grid treats both the same. It also says nothing about whether your week has room for everything you’ve put in the “do” and “schedule” boxes. So a box can be full in a way the grid can’t show you. Six urgent and important tasks look the same on the page whether they add up to five hours or fifty.
The fix is plain arithmetic. Put a rough estimate of hours beside each item. Next to that, write down the hours you actually have free this week once meetings and the work that always turns up are taken out. If the total is bigger than the hours, something has to be named as not happening this week. That’s a decision, which is different from a sort, and it’s one the plain grid lets you avoid. Keep the estimates rough, though. Half an hour, half a day, a day and a week are precise enough. Spending longer on the estimates than on the work is its own trap. Larger pieces of work may need more than a rough estimate. You may need to rank them by impact against effort, or score them with RICE or ICE. Our roundup of prioritization frameworks covers those, along with MoSCoW, which fits work to a fixed amount of capacity. One more cost to keep in mind: estimating forty items is tiring, and decision fatigue is a real reason to estimate only the items you’re unsure about.
It sorts one person’s list, but managers arbitrate between people
Covey presents the matrix as a tool for personal management, so it works best on a list that belongs to one person. But a manager’s list rarely does. It mixes their own work with what everyone else wants from them, and that breaks the grid in three places.
The first problem is that the requester often sets the urgency. Covey makes this point himself. He observes that much of what feels urgent in a busy day gets its urgency from other people’s priorities and expectations. So when your director writes “need this by Thursday”, the task arrives already labeled. You sort it as urgent because someone said it was.
The second is the “delegate” box. It assumes there’s someone with spare time to hand the task to, but a manager’s team is usually as full as the manager. Delegation is a real skill with real choices in it, such as how much authority goes with the task. Our guide to the seven levels of delegation covers those choices. Even so, a box that says “delegate” doesn’t tell you who has the time.
The third causes the most trouble. The grid has no place for two senior people who want opposite things from the same week. So both requests land in the urgent and important box. The manager quietly picks one, and the person who loses finds out when their deadline passes.
The repair is to turn the first failure’s question outward. When a request arrives marked urgent, ask the person who sent it what happens if it waits. Also ask until when it can wait. Our MoSCoW prioritization guide works through a similar test for a release, where two stakeholders both call their item a must. Here the requests arrive one at a time across a week, so you ask as each one lands. If two requests collide, put them side by side in front of both people in one message or one short conversation. Say what each would take, what you know about the cost of each waiting, and what you propose. The order is still your decision, but now you’ve made it openly. That means the person who gets the later slot hears it from you rather than discovering it on Thursday afternoon.
This costs you something, because it’s slower than sorting privately and it exposes you. Someone will hear a “not yet” in front of a peer, and some senior people don’t enjoy being asked what their deadline is based on. That’s why this repair is less about the grid than about people skills: asking the question without sounding as if you’re refusing the work.
The “schedule” box never gets a date
The important but not urgent box is the one Covey treats as the most valuable, yet it’s the one that never feels due. So a task placed there has been parked, not planned. The trouble is that the grid is a snapshot of how things look on Monday morning. Because it doesn’t carry anything forward, the same items sit in the same box week after week. Each time they look as reasonable as they did the week before.
The fix is to give every item in that box a day, an owner and a block on the calendar. You also re-sort the whole list at a fixed time each week, say Monday at nine. A block on the calendar only helps if you can keep it, and our guide to protecting deep work covers how to defend one when your calendar isn’t fully yours. The cost is a fuller calendar with less slack in it. If a block keeps getting moved week after week, that tells you something too. It may mean the item wasn’t as important as you labeled it, which sends you back to the second failure.
One Monday list, sorted twice
Rosalind and everyone in her week are composites, made up for this post rather than drawn from real people. But the week itself is an ordinary one for a manager in her job. She manages a customer-service team for a company that repairs household appliances. The team is nine agents and a team lead, Edwin. She has about 12 hours this week that are hers to plan. That’s what is left after her standing meetings, time covering the phones and the escalations that come to her by default. She has three outcomes for the quarter. The first is fewer customers having to call back about the same repair. The second is the two agents who joined last month taking calls on their own by the end of the quarter. The third is a winter shift roster agreed before December, when call volumes rise.
The first pass
At half past eight on Monday she writes everything down and sorts it the standard way, which takes about ten minutes. The whole list is below, with her rough hours and the box each item went into. It also shows where each item ended up after the second pass.
| Item (rough hours) | Who asked | First pass | After the second pass |
|---|---|---|---|
| Quarterly call-quality report in the new format (4) | Alistair, regional director | Do now | One-page summary by Friday noon (1.5); full report next month |
| Complaints review for the biggest client’s meeting on Thursday (4) | Joanna, key accounts manager | Do now | Done in full by Thursday morning |
| Reply to a customer escalation (0.5) | Complaints inbox | Do now | Do now |
| Approve three holiday requests (0.5) | Two agents | Do now | Do now |
| Sign off timesheets before payroll closes Wednesday (1) | Finance | Do now | Do now |
| Desk plan for the floor refit (2) | Esther, facilities | Do now | Moved to next Tuesday |
| Rewrite the script for the most common repeat-call fault (3) | Rosalind | Schedule | Friday 9 to 12, Rosalind, with Edwin’s call notes |
| Sign-off sessions for the two new agents (3) | Rosalind | Schedule | Tuesday (Rosalind) and Wednesday (Edwin) |
| Draft the winter shift roster (3) | Rosalind | Schedule | Not happening this week; booked for next Monday |
| Revamp the team newsletter (2) | Rosalind | Schedule | Dropped for this quarter |
| Demo of a call-recording tool, “offer ends Friday” (1) | Vendor | Delegate | Dropped |
| Tidy the shared drive (1) | Rosalind | Drop | Drop |
The first pass looks sensible, and on its own terms it is. The do-now box holds six things, and each has a real person waiting. Alistair, her regional director, wants the quarterly call-quality report in a new format by Thursday for a board pack. Joanna manages the company’s largest client, a property management firm that books repairs for a few thousand rental homes. She wants last month’s complaints reviewed before she meets that client on Thursday morning. Esther in facilities asked for a desk plan for the floor refit “by end of day”. Those six come to 12 hours, which is every hour Rosalind has. That means her own four items in the schedule box, which add up to another 11, get no time at all. The vendor demo sits in the delegate box, although everyone she could hand it to is on the phones all week.
Nothing in that grid is wrong, exactly. But it tells her she’ll spend the whole week on other people’s deadlines and none of it on the three things she’s measured on.
The second pass
She goes back to the urgent column first and writes beside each item what happens if it waits. The holiday requests and timesheets keep their place, because two agents are waiting to book flights and payroll closes on Wednesday whatever she thinks of it. The escalation stays too, because that customer has already been promised a callback twice. The desk plan is where her urgent label drops. The refit isn’t until January, so she writes that nothing happens if the plan is done next week. She moves it to next Tuesday. She writes that line herself from what she knows about the refit. She doesn’t ask Esther, and that matters later.
The vendor demo goes next. Its urgency is a discount that ends on Friday, but she has no budget for a new recording tool this quarter at any price. It’s the kind of manufactured expiry the urgency experiments used, so she drops it.
Next comes the outcomes test. The newsletter revamp had gone into the schedule box because the team likes the newsletter and she’d promised to freshen it up. But when she has to name which of her three outcomes it serves, she can’t. So she drops it for the quarter, without much confidence that it’s the right call.
The two Thursday requests are the hard part. Together they need eight hours out of the twelve she has for the whole week, and both come from people more senior than her. On the first pass she would have quietly done Alistair’s report first because he’s her boss, and then stayed late on Joanna’s review. This time she sends one message to both of them instead. She sets out what each piece of work would take and says she can’t finish both in full by Thursday. She also gives what she knows about the cost of each one waiting. Joanna’s client meeting can’t move, but the board pack doesn’t go to the board until the following Monday. So she proposes the complaints review in full, plus a one-page summary of the call-quality figures in the new format by Friday noon. The full report would follow next month. Alistair replies within the hour and agrees. He adds that next quarter he wants the full report on time. He also says he’d rather have heard about the clash from Rosalind before it reached him alongside Joanna.
Now the arithmetic works. The fixed items come to 7.5 hours. Two of them are the complaints review and the report summary. The escalation, the holidays and the timesheets make up the rest. That leaves 4.5 hours for nine hours of her own work. Edwin agrees to run one of the two sign-off sessions for the new agents. That brings the item down to an hour and a half for Rosalind on Tuesday afternoon, with Edwin owning Wednesday’s. The script rewrite gets Friday from nine to twelve, with Rosalind as owner. Edwin will also pull the notes from ten repeat calls by Thursday so she isn’t starting from nothing. That uses the last three hours. The winter shift roster doesn’t fit, so she names it as not happening this week and books Monday morning of next week for it. She also tells the team at the Monday huddle so nobody’s surprised.
The second pass took her about fifty minutes, against ten for the first. She also spent an hour waiting on Alistair’s reply before she could fix the plan.
How the week went
On Thursday afternoon Esther emailed. The refit contractor would only hold its January slot until Friday, and it needed the desk layout to confirm. Rosalind’s note had said nothing would happen if the plan waited a week. That had been a guess, and it was wrong because she’d never asked the one person who knew. So she did the plan first thing on Friday, using two hours she didn’t have. Friday’s three-hour block on the script shrank to one as a result. This is the gap the urgency section flagged. The workers in the experiment were only reminded of payoffs they already knew. But Rosalind had to estimate a cost, and she estimated it without asking the requester. Asking was the fix in the fourth failure.
The rest went roughly to plan. Joanna went into her client meeting with the full complaints review. Alistair had his summary at eleven on Friday. One of the new agents was signed off to take calls alone, but the other needs another week on billing questions. The roster is booked for Monday.
Two things are still open. The first is the script rewrite, which is the item tied most directly to her first outcome, and it got one hour out of three. It’s back on next Friday’s calendar. If it moves a second time, the fifth failure says that tells her something about how important it really was. She isn’t ready to accept that. The second is the newsletter. Two agents asked where it had gone at Friday’s huddle, and she hasn’t decided whether dropping it cost more than it looked.
When is the plain matrix the better call?
What follows is editorial judgment too, and it’s about the limits of everything above.
The repairs cost time, but speed is the whole value of the matrix. Rosalind’s second pass took five times as long as her first. So on a short list with goals you already know and deadlines that are real, the repairs are overkill. The plain ten-minute sort is the right tool there.
The second failure may not be a flaw in the matrix at all, because Covey’s chapter assumes you’ve already done the work of deciding what matters before you reach the grid. If you’ve done that work, “important” already has a yardstick. In that case the failure is really a matter of using the grid without the step that was meant to come first. That’s a fair reading, and it would make one of our five failures a case of misuse rather than a design problem.
The evidence for the first failure is also narrower than it might sound. The mere urgency experiments used small stakes, short tasks and deadlines invented for the purpose. So they show a tendency, not that managers misjudge urgency at any measurable rate. The research done for this post found no study that tests the matrix against any alternative in real workplaces, and none that tests these repairs either. That means whatever you try here, you’re trying it on reasoning rather than on proof.
There are also days when the plain grid is exactly what you need. On a day of truly competing deadlines there’s nothing to estimate, because everything is due and you just need an order. Sometimes you’re overwhelmed and need ten minutes of structure before you can think at all. The grid gives you that, and the repairs would get in the way. If the list belongs to you alone, the manager failure doesn’t apply. Many choices also don’t deserve a method in the first place. If a decision is easy to undo, our guide to reversible decisions explains why you should usually just make it and move on.
The repairs can also be overdone. A manager who writes a consequence line for every email has rebuilt the problem in a new form, just with more paperwork attached.
What is the best alternative to the Eisenhower Matrix?
The best Eisenhower Matrix alternative for most managers is a division of work rather than a replacement. The matrix keeps the job it’s good at, and other tools take the jobs it isn’t. The table below shows the split.
| The job | What to use | Why |
|---|---|---|
| Sort today’s crowded list | The Eisenhower Matrix, as it is | Fast, needs no scoring, and separates what has to happen today from what can wait |
| Decide what deserves your hours this week | The matrix plus the repairs above | Adds the cost of waiting, the outcomes behind “important”, the hours you have, and a date for the “schedule” box |
| Rank competing projects or requests at scale | A scoring or must-have method | Compares many items on the same terms, which the four boxes can’t |
Our prioritization frameworks roundup compares RICE, ICE and Kano for the third job. It also makes the point that whatever framework you pick, the hard part is still saying no. Our guide to running MoSCoW covers sorting a fixed amount of work into what must ship and what waits.
If you only try one thing from this post, make it the first repair. Tomorrow morning, before you sort, pick the three items that feel most urgent and write beside each one what actually happens if it waits a day. If you can’t say, ask the person who sent it.
That question is the part most people skip, because asking someone senior what their deadline is based on can feel like refusing the work. So you might want help working out what waiting really costs before your reply goes out. You might also want help wording the message that puts two competing requests in front of the people who made them. You can talk it through with Merlin in Slack or Microsoft Teams.