How to prioritise a long backlog
The short answer
Do not sort it. Sorting two hundred items produces a ranking that is stale within a week and took an afternoon to build.
A backlog is not a plan and does not need a total order. It needs three things:
- A top, which is short — the ten or so items that are genuinely next.
- A bin, used regularly and without ceremony.
- Everything else left unsorted, where it costs nothing.
The mistake is reading the backlog as a queue to be ordered. It is a pool to be chosen from, and the choosing happens at the day.
Why a full ranking is wasted work
Most of it will never be done. That is not a failure of the backlog; it is what a backlog is. Ranking items 60 through 200 orders things that will be deleted or overtaken before they are reached.
The order is wrong by next week. Priorities that depend on the world change when the world does. A ranking is a snapshot presented as a plan.
It feels like progress and produces none. An afternoon of sorting ends with the same amount of work done as an afternoon of nothing, and a stronger feeling of having been productive — which is the more expensive of the two outcomes.
Ranking cost scales badly. Comparing two hundred items pairwise is not something anyone actually does, so what really happens is a rough sort by recency and guilt.
What to do instead, in twenty minutes
Pull the top ten to the surface. Not ranked among themselves — just separated. These are what could plausibly be next. Everything else stays where it is.
Delete without ceremony. Anything older than three months that has never been close to the top is not going to be done. Deleting it is not giving up; it is removing something that costs attention every time the list is read and returns nothing. A backlog nobody deletes from becomes a record of intentions rather than a source of work.
Merge the duplicates. Long backlogs always contain the same thing written four ways at different times. Merging them is usually the single biggest reduction available.
Tag only what is blocked, and on whom. Not categories, not effort scores — just blocked or not. Blocked items cannot be next however important they are, and knowing that at a glance saves picking one by mistake.
That is the whole review. Twenty minutes, once every week or two.
Why a long backlog is not itself the problem
Worth being precise, because guilt about backlog size drives a lot of wasted sorting.
By Little's Law, average completion time is proportional to work in progress — things started and unfinished. A captured item you have not started is not in progress and costs you nothing in throughput. A hundred unstarted items do not slow anything down. The arithmetic.
What slows you down is how many things are simultaneously open. So the discipline that matters is at the day — a small number, chosen and closed — not at the backlog, which can be as long as it needs to be. More on list length.
The exception: a backlog gets expensive when reading it costs something. If opening it produces dread rather than options, that is the signal to delete, not to sort.
Choosing from it
Once a day, take a small number — three is practical — from the top ten. The question is not "what is most important overall?" but "what most needs to be true by this evening?", which is a question the day can answer and the backlog cannot.
Anything that keeps reaching the top ten and never gets chosen is telling you something specific, and it is usually not a discipline problem.
Where to go next
- Little's Law, applied to a working day — why the backlog length is not the constraint
- Is a long to-do list bad? — capture and planning as separate surfaces
- How to prioritise your tasks — the three conditions
Using this in Decide One
Decide One keeps everything captured separate from what you committed to today, because they are different objects with different rules. The pool can be long. The day is bounded, and the bound is enforced rather than suggested.
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.