The Journal

The small job that costs more to remember than to do

A small household job is cheap to do and expensive to carry. The difficulty is not discipline. It is the distance between the moment a job is noticed and the moment there is time to do it.

---

The arithmetic of a small job

Every house has a list it never wrote down. The hall bulb. The parcel to return. The form that needs a signature. Most are the work of a few minutes.

What they cost is not the doing. It is the carrying. A job that has been noticed and not done must be remembered, or written down and then found again. It is re-decided every time it is walked past. It is mentioned to the other adult in the house, then mentioned again, then asked about. Set against a few minutes of work, that accounting is absurd.

The best-known statement of this is David Allen's two minute rule, from Getting Things Done, published in 2001 and rewritten in 2015. It covers any action that, in Allen's words, "can be done in two minutes", on the reasoning that organising it and reviewing it later costs more than finishing it the first time you notice it. He meant it narrowly. He describes it as a question asked while working through an inbox, a device for forcing a decision about what the next step actually is, rather than a rule for every thought that arrives during the day. Nothing in the published record tests it: not the two-minute threshold, and not doing a job now against capturing it and batching it later. It is a consultant's rule of thumb, and it is more credible stated as one. The rule of the same name in James Clear's Atomic Habits is a different idea, about shrinking a new habit to its first two minutes.

Why small tasks pile up, and why it is not laziness

The reason a five-minute job survives six weeks is structural. The moment of noticing is almost never a moment of freedom. A tap drips while somebody is carrying a sleeping child. The noticing happens with both hands full, and by the time both hands are empty the job has gone from mind and the house looks normal again.

The closest measurement comes from the workplace. Gloria Mark and colleagues shadowed 24 information workers and found they spent an average of 11 minutes 4 seconds on a stretch of work before switching, that 57 per cent of those stretches were interrupted, and that work resumed the same day came back after an average of 25 minutes 26 seconds, with more than two other activities in between. A later lab study led by Mark, with 48 participants, found interrupted work was finished faster than uninterrupted work, not slower, but with significantly more stress, frustration, time pressure and effort. The widely quoted figure of 23 minutes and 15 seconds to refocus has no primary source; it traces to interviews and press quotations rather than to any published result.

A home is a harder case than an office: several rooms, several people, no inbox. Willpower is not the missing ingredient, and advice that starts from that assumption will fail.

What the evidence actually supports

The strongest result here is not about doing jobs quickly, but about getting them out of your head and into something else. Sam Gilbert ran four experiments with 1,196 participants and found that setting an external reminder significantly improved accuracy on delayed intentions, and that where offloading was not allowed, an interruption left people less accurate at carrying out the intention they were holding. Capturing an intention somewhere outside memory is the part of this subject with real support behind it.

Several familiar claims do not survive contact with the sources. The idea that unfinished tasks are remembered better, usually called the Zeigarnik effect, failed a 2025 meta-analysis of 59 publications, which put the ratio of interrupted to completed recall at 0.99. What held up was a different tendency, to resume an interrupted task, at about 67 per cent. Planning fares better: a meta-analysis of 642 tests of if-then plans found real effects, though its authors report substantial publication bias and warn that the headline figure should not be read at face value. Whether time management works has one decent answer: across 158 studies and 53,957 people it relates moderately to performance and more strongly to wellbeing, though the data are mostly cross-sectional, and the review cannot say which technique does what, because the studies rarely record what was taught.

Two findings cut against doing every small thing the second it appears. Across nine experiments with 257 people, participants took on extra physical effort in order to start a subtask sooner, which the authors put down to unloading working memory. Separately, across five experiments, people chose tasks with objectively lower payoffs when those tasks merely looked urgent. Always doing the smallest visible thing turns attention into a hopper for whatever is nearest.

The household version: deciding, then deciding again

In a shared home the carrying has a second layer. Somebody has to notice the job, work out the options, decide, and then check it was done. Allison Daminger, drawing on 70 interviews with members of 35 couples, named those four steps cognitive labour and found them unevenly divided: the women in her different-gender couples did more of it overall, and more of the anticipating and the monitoring in particular, while the deciding was shared about equally. A cross-sectional survey of 322 mothers found they reported 72.57 per cent of that work, with associations to higher stress and lower relationship quality. A 2023 review of 31 studies reached the same descriptive conclusion, and noted that no validated scale for measuring mental labour exists at all. The load is real and repeatedly measured. What causes what is not established, and nothing here says a planner changes it.

What this adds is negotiation. A job with no owner is re-decided by two people instead of one, which is how a job of a few minutes acquires a fortnight of overhead.

How to stop putting off small jobs

The honest advice is not to try harder. It is to shorten the gap between noticing and doing, which can be worked on from either end. Only the first of the suggestions below has research behind it. The rest are this paper's opinion, offered as such.

Capture at the point of noticing, in one place, in the fewest taps that will hold the thought. This is the part with evidence behind it. Write the physical action rather than the topic: "hall bulb, bayonet, two of them" is a job, while "hall" is another decision deferred.

Give the job an owner once, rather than every time it comes up. The negotiation, not the screwdriver, is the expensive part.

Tie it to a moment rather than a date. Most small jobs have no real deadline, so a due date is a fiction that goes red and then gets ignored. A prompt arriving when there might be a gap fits these jobs better.

Some should wait, and that is fine. The aim is not to obey a stopwatch, but to make sure a job sits somewhere reliable and comes back when doing it is possible.

What Barriers does about it

Barriers is a household planner from IV Technologies Ltd, a UK company. It holds tasks, money and receipts for the adults sharing a home, and one part of it is built for this gap.

Any open task can be tagged by hand as a quick task, either from a row at the bottom of its card or from a toggle in the sheet where it is written down. Nothing is estimated: the app has no duration field and makes no judgement about how long a job will take. The tag is the user's own opinion, recorded at the moment of noticing.

Tagged tasks are then brought up at times the app picks rather than at a deadline. There are four notification slots a day in total, not four per task: one inside each of four two-hour windows between 8:30 and 19:30, at a random minute. The slots rotate through the tagged tasks in deadline order, so a long list is shared out rather than one job speaking every time. If the work-hours window in settings has been switched on, any slot falling inside it is skipped rather than moved.

The prompt reads "Got five minutes?" and names the task. Pressing and holding it reveals three buttons: Done, Tomorrow and Complete Now. There is no snooze and no decline. The nearest thing is Tomorrow, which moves the job to tomorrow at its usual time, or leaving the notification where it is. Done and Tomorrow both work from the lock screen without opening the app. Tapping the notification, or Complete Now, opens the app and starts a five-minute countdown: a Live Activity on the iPhone Lock Screen and Dynamic Island, a strip inside the app, a countdown notification on Android. When the time is up the Lock Screen offers Completed and Not completed, the second of which only dismisses the timer and changes nothing about the task; elsewhere a plain "Five minutes are up" notification stands in. The same countdown can be started deliberately from a task's own menu.

The five minutes are the length of the timer, not a claim about the job. A task leaves the pool when it is completed or parked, and tapping the tag again stops it, which is how the nudges are turned off: one task at a time, or through the phone's own notification settings. These are local notifications with no server involved, armed a week at a time whenever the app runs, so an app left closed for longer than that goes quiet until it is opened again. In a shared household a quick-task tag applies to your own copy of a job, not to anyone else's.

None of that is a cure for anything, and it does not need to be. It is a way of making sure a job is written down once, and comes back when there might be five minutes free.

---