What to do with work that repeats
The short answer
Recurring work should not be re-decided every time it comes round. Deciding is the expensive part, and for work that repeats you can pay it once.
Scheduling theory gives periodic work a class of its own, and its standard answer is rate-monotonic scheduling — assign a fixed priority by how often something recurs, more frequent meaning higher priority, and then stop deliberating. Liu and Layland showed in 1973 that this is optimal among fixed-priority schemes for periodic tasks.
The transferable idea is not the formula. It is that routine work should have a settled position, decided once, so that each occurrence costs no decision at all.
Why re-deciding is the cost
A recurring task that gets weighed on its merits every week costs you the deliberation every week — and deliberation is a switch, paid in reload rather than in minutes. Fifty-two small decisions a year about whether to do the weekly review is more expensive than the review.
It also produces a predictable failure: the recurring item loses every time it is weighed against something urgent, because it will run again next week and the urgent thing will not. Each individual deferral is reasonable. The cumulative effect is that the routine never runs — starvation, with a weekly cadence.
Settling the position once removes both costs.
Giving each one a fixed position
More frequent, more fixed. Daily things should be the least negotiable and attached to something that already happens — the close of the day, the first coffee. A habit with no trigger is an intention.
Less frequent, more explicitly protected. A monthly task has eleven chances to be skipped before anyone notices, so it needs a real slot rather than good intentions.
Same slot every time. The point is that the question stops being asked. A routine that moves is a routine being re-decided, which is the thing you were avoiding.
The three questions worth asking, but only occasionally
Not every occurrence — every quarter or so:
Does this still produce anything? Recurring work outlives its reason more reliably than any other kind, because nothing ever prompts a review. The report nobody reads is the archetype, and it survives because cancelling it requires someone to act while continuing it requires nobody to.
Is the frequency right? Most recurring work is set at whatever cadence seemed natural at the time. Weekly things that could be fortnightly are extremely common; the reverse is rare.
Can it be batched or automated away? Several small recurring items of the same kind usually merge into one block, which costs one load-in instead of several. Where batching helps and where it quietly hurts.
Deleting a recurring task is worth more than deleting a one-off, because you are removing every future occurrence at once.
What not to make recurring
Anything with a real deadline. Deadlines and recurrence are different disciplines, and mixing them means the deadline gets the routine's flexibility — which is exactly the wrong way round.
Anything you have skipped three times running. That is not a recurring task, it is a recurring reminder that you are not doing something. Either protect it properly or delete it; leaving it to reappear costs attention every cycle and returns nothing.
Reviews, on bad weeks. The review is worth protecting but not worth enforcing. A routine that becomes a thing to feel bad about is a routine that gets abandoned, and then the useful part goes with it.
The honest limit
Rate-monotonic scheduling is a result about periodic tasks in computing systems with known periods and known durations. A working week has neither. The analogy holds for the thing that matters — a fixed position removes a recurring decision — and it should not be pushed further than that. It does not tell you how often to do your weekly review.
Where to go next
- Is there a best prioritisation method? — where rate-monotonic sits among the other disciplines
- Should you batch similar tasks? — merging small recurring work
- How to prioritise your tasks — the three conditions
Using this in Decide One
Decide One holds one day at a time and does not manage recurring schedules — that is a deliberate limit rather than a missing feature. What it gives routine work is the same thing it gives everything else: a real duration, so a day full of routines is visibly a day full of routines.
A line you did not finish is unfinished. It is not marked against you, and a day you did not open is not counted.
Open Decide One — free to use, with no account. Nothing you write leaves your device.
Decide One is a priority instrument. Bring the day into view, choose a method, give the first thing real time.