ManageTime

Article

Work nobody is chasing you for: the task with no deadline at all

To get an undated task done, give it a slot in this week rather than a date, and put one other person on the other end of it. A date you set privately is the weakest form of deadline, and a 2026 replication in Psychological Science found that changing deadlines moved performance far less than the classic study reported. This page covers why undated work stalls in a way dated work does not, what the deadline research still supports after that replication, and which of the site methods survive the absence of a due date.

Why do you keep putting off work nobody is asking about?

Fix the onboarding doc. It has sat on your list since June, nobody has mentioned it once, and nothing whatsoever happens on the day you skip it again. Every other line on that list carries a date, or a person, or both. This one carries neither, so it competes for your Tuesday against work that will be noticed by Wednesday, and it loses that contest quietly every single week.

The loss never looks like a loss, which is the part worth naming. The procrastination pillar sets a clean test for telling a stall from a decision: if you can name what the task lost to, it was ranked, and if it simply failed to happen, it stalled. Undated work defeats that test. You can always name what it lost to, because everything carrying a date outranked it, so month after month the stall files itself under sensible prioritizing and never comes up for review.

Difficulty is not the variable here.

Somebody has measured the nearest version of your situation, and the numbers are not flattering. In 2022 Steel, Taras, Ponak and Kammeyer-Mueller published a study in Frontiers in Psychology on what they call slippery deadlines, where the due date is ill defined and the person doing the work can extend it themselves. Their setting was grievance arbitration: three years of observed performance data from Canadian arbitration cases, plus a survey of leading arbitrators. Trait procrastination predicted observed delay about as strongly as the complexity of the case did. Arbitrators one standard deviation above the average on procrastination took roughly 83 days to write their decisions, against 26 days for those one standard deviation below.

Read the setting before you borrow the numbers. Those are trained professionals, facing real consequences for delay, working cases that at least carry a due date they are able to move. Your onboarding doc carries none. The study does not measure your task, and what it does establish is that the power to move a date is on its own enough to stretch the same work by two months.

What sits in your undated column matters more than its position suggests. The documentation nobody reads until the week you leave, the process fix that would save an hour every Monday, the reference file that would answer the question you get asked twice a month: none of it has a requester, and all of it compounds. Dated work arrives with a reason to happen this week. Undated work has to be given one.

Does setting your own deadline actually work?

Invent a date, says page one of the search results. Put it in the calendar, treat it as real, tell yourself the doc is due Friday. That is decent advice resting on a thinner evidence base than its confidence suggests, and the base got thinner in 2026.

Start with where the confidence came from. In 2002 Ariely and Wertenbroch published a paper in Psychological Science asking three things about self-imposed deadlines: whether people will set costly ones for themselves, whether those deadlines improve performance, and whether people set them well. The answers were yes, yes, and no. People recognize their own self-control problems and will pay to constrain them, the deadlines they set do help, and the dates they pick are not the dates that would help most. That third finding gets quoted least and matters most to you, because the date you would invent for the onboarding doc is precisely the kind the study found people choose badly.

In 2026 somebody ran the experiment again.

Hyndman and Bisin published a replication of that paper's second study, also in Psychological Science, using adult participants enrolled at a large public university in the United States. It did not replicate. Changes in the deadlines had a negligible effect on the three performance metrics the original used, and on several of its survey measures too. The evenly spaced deadlines imposed by the experimenters, which in the original stood out as the most effective arrangement, did not stand out here.

Take that at its actual size. One failed replication does not prove that deadlines do nothing, and what got tested was the arrangement of deadlines rather than the value of having one at all. What it removes is the confident version of the claim. The specific result showing that a well-placed deadline reliably lifts output was the strongest published support for inventing a date, and it did not hold on a second run, so anyone still selling you the invented date as the mechanism is selling it without that.

The samples are the other half of the caution. Both the original and the replication ran on students doing set tasks for a course, and the one workplace study on this page ran on arbitrators. Nobody has tested an invented date against a backlog like yours, with a manager attached, a calendar you only half control, and no consequence at all for letting the date slide. Treat the direction as the part that transfers and leave the numbers where they were measured.

So the date is not the lever. What the procrastination pillar already says about artificial deadlines points the same way: a deadline only you know about is one you are free to move, and on a bad Tuesday you will move it. For work that never had a date in the first place, that is the whole design problem: your commitment has to sit somewhere other than in a date you wrote for yourself.

How do you get an undated task started?

Take the line you have carried longest. Three moves, in this order, and none of them is picking a Friday.

Give it a size first. Undated work is usually unshaped work too, and the pillar names that pairing plainly: fix the reporting and sort out the handover are headings rather than tasks, with nothing in them you can open for twenty minutes. Timeboxing imposes the shape by the clock. You fix how long the sitting runs before you begin, and a box that ends at a set time cannot be entered without a definition of enough. Two hours, once, to produce a draft outline is a task. Fix the onboarding doc is not.

Then give it a slot rather than a date. A slot is a Tuesday morning inside a week you can already see, and it competes on equal terms with the dated work around it because it occupies the same hours those meetings do. A date three Fridays out competes with nothing this week. That difference is why an invented deadline can sit in your calendar for a month while the task it names never begins.

Then hand one end to somebody.

The evidence above makes this move compulsory rather than nice to have. Parkinson's Law is the site's method for manufacturing pressure, and its page is specific about the second half most people drop: promise the delivery to a person, or book the next commitment immediately after your earlier date so that something real sits behind it. Send the outline to the colleague who keeps asking the question your doc would answer, on the day your slot ends. Ariely and Wertenbroch found the deadlines people set for themselves helped less than some of the ones imposed on them, and the replication was unimpressed by the imposed schedule as well. No study is needed for the plainer point: a date another person is expecting is harder to move on a Tuesday than a date only you can see.

Order is the alternative when no person is available. The Ivy Lee Method asks for six tasks written the evening before, ranked one to six, worked in sequence, with no permission to start the second before the first is finished. Put your undated task at two or three. Its problem was never a low rank; it was that the task never entered the ranking at all, and a written order enters it by force.

One check before any of this. Ask what breaks if the task never happens, and answer honestly rather than defensively. Some work is undated because nobody needs it, and a size, a slot and a colleague will get that work delivered on schedule, which is the wrong outcome reached efficiently. The pillar makes the same point about every method it lists: structure is good at making work start, and it says nothing about whether the work was worth starting. If nothing breaks, delete the line instead of booking it.

How does your working style change the answer?

Which of the three moves you skip depends on how you already work. Architect, Sprinter, Visionary and Improviser are this site's four names for that, set out in the working-style self-check, and pairing them with undated work is editorial judgment rather than a research finding.

Architects have the least trouble with the slot and the most with the size. An undated task gets a well-formed plan and a place in the week, then stalls at the point where somebody has to say what a finished version looks like, because that answer is open and a plan can be improved while it stays open. Write the definition of enough badly, on purpose. You can revise it in the slot.

Sprinters need the size before anything else, for the reason the pressure question exposes elsewhere on this site: a burst that meets a task twice its assumed size is not a plan. Undated work is the worst offender there, because nothing has ever forced anybody to estimate it. One hour, on the clock, to produce the outline only.

Visionaries usually hold the finished version vividly and the start date not at all, which is the combination undated work punishes hardest. The outcome is clear, the appetite is real, and no Tuesday is attached to either. Skip to the third move: name the person who gets your first version, and tell them when.

Improvisers get the slot booked over. A Tuesday morning that survives contact with your week is not something you can count on, so tie the task to an event rather than a clock, and keep the first move small enough to survive a day going sideways. The Ivy Lee Method's ranked list does that job without depending on a calendar holding its shape.

So take the oldest undated line on your list today. Write down the hours you think it needs, book one slot inside the next seven days, and send one message telling somebody what they will get at the end of it. If the slot survives and the message goes out, you will know within a week whether the missing date was ever the problem. If neither happens twice running, the question you owe that task is a ranking one, and the prioritization methods answer it rather than this page.

Frequently asked questions

How do you get tasks done when there is no deadline?

Give the task a size and a slot in the week you can already see, then put another person on the receiving end of it. A date you invent for undated work sits three Fridays out and competes with nothing on Tuesday. A booked hour competes with the meetings around it, because it occupies the same time they do. The person is what stops you moving it quietly, and it is the part most people leave out.

Does setting your own deadline actually work?

Less reliably than the advice suggests, and the strongest study behind that advice did not survive a replication. Ariely and Wertenbroch reported in 2002 that people will set costly deadlines for themselves, that those deadlines improve performance, and that people choose the dates badly. A 2026 replication of that paper's second study, in the same journal, found changes in the deadlines had a negligible effect on the three performance metrics used originally. Set a date anyway if you like, but do not build the week on it.

Why do you keep putting off work nobody is asking about?

An undated task loses every day to work that carries a date, and the loss never looks like a decision. The test this site uses for a stall is whether you can name what the task lost to. Undated work always hands you an answer, since everything with a due date outranked it, so the stall reads as sensible prioritizing for months. The work sitting in that column tends to be the work that compounds.

Which method suits work with no due date?

Timeboxing gives it a size, and the Ivy Lee Method gives it a place in the order when no date can. A timebox will not let you begin without a definition of enough, which is exactly what an undated heading lacks. A ranked list of six tasks worked in sequence puts the task into a queue it would otherwise never enter. Parkinson's Law covers the invented deadline, and its page names the half that makes it hold: tell somebody, or book the next commitment right behind it.

Sources