When nobody has time to document
The objection is usually correct. An approach that starts by asking for a documentation project loses, and should — what works is making each document small enough that the objection stops applying.
How to get procedures written in an organisation where the objection that nobody has time is a true statement about the current week.
The objection is usually correct
"We don't have time to document" is treated as an excuse and it is generally a true statement about the current week. The people who know the processes are the people the work depends on, and the hour taken from them is a real hour that something else was going to use.
So an approach that starts by asking for a documentation project loses, and it should. What works is the opposite: making each document small enough that the objection stops applying to it.
Stop trying to document the process
The instinct is to capture a whole area of work — onboarding, month end, the release. That is a week, it needs a quiet week, and there are no quiet weeks. The project stalls at 60% and the half-finished document is worse than none, because nobody knows which parts are current.
Document one task instead. The narrowest unit somebody could hand to another person and walk away from. It fits in twenty minutes, it is finished the same day, and a finished narrow document beats a stalled broad one on every measure that matters.
The unit test: if you cannot name who does it and when it happens, it is a process rather than a task and it needs breaking down further. The distinction between the two is what stops this going wrong.
Capture while the work happens, not afterwards
Writing a procedure from memory is slow and inaccurate — it takes an hour, it is done in the wrong order, and it silently omits the steps the person no longer notices doing.
Capturing during the work costs almost nothing. Whoever is doing the task keeps a note open and writes a line per step as they go, in whatever words arrive. Five minutes on top of work that was happening anyway, and no separate session to schedule.
The output is not a document. It is a correct list of steps in the right order, which is the expensive part. Turning that into a document is twenty minutes of tidying that anyone can do — and it does not have to be the person whose time is scarce.
Getting a procedure out of the person who holds it covers the version where somebody else does the capturing.
Write it the next time it happens, not now
The rare, consequential tasks — the annual filing, the quarterly close, the renewal — are the ones most worth documenting and the hardest to schedule, because there is no natural moment.
The natural moment is the next time it happens. Put a line in the calendar entry that says to capture it, and accept that the document does not exist until then. That is a wait of weeks or months and it is still faster than a reconstruction that nobody starts.
The same logic applies to anything done rarely. If it happens twice a year, you get two chances a year, and taking one of them is the whole strategy.
Let the first version be bad
A large amount of the time cost is the standard people are holding themselves to. A procedure written to look like a manual takes four times as long as one written to be followed, and it is not four times as useful.
The first version can be a numbered list with no purpose statement, no revision block and no formatting. It is already worth more than the nothing it replaced. The control block, the numbering and the review trigger can be added later — and it is worth adding them, but not at the cost of the document existing.
The one thing worth doing properly from the start is having somebody else follow it once without help. That is the step that finds the missing steps, and it is the step that gets skipped.
An order that fits a real fortnight
Pick the single task where one person is the only one who can do it and the work stops if they are away. Not the most complicated one — the one with the worst exposure.
The next time that task runs, whoever does it keeps a note open and writes a line per step.
Somebody else — not necessarily them — spends twenty minutes turning the lines into a page.
Somebody who has never done the task follows the page, without help, and the gaps get filled in.
Then stop. One document a month, current and used, beats a library that was written in one push and is stale by spring.
Every template here is free, blank and worked, and needs no account. The SOP template is the two-page format this post assumes, and how to write one covers the tidying step.
Editor's notes
Written by Free Biz Docs about the document above — not reader submissions.
The post concedes the objection in its first line
"We don't have time" is treated everywhere else as an excuse to be overcome. It is generally a true statement about the current week, and a page that argues with it loses the reader immediately. Conceding it is what makes the rest of the argument available: the fix is not more time, it is a smaller unit.
"Let the first version be bad" sits badly next to document control
This site's whole visual argument is control — numbering, revision, ownership, a block at the top of every page. Telling somebody to skip all of it for the first draft is a genuine tension rather than a rhetorical one.
It is resolved in one direction: a controlled document that does not exist controls nothing. Add the block at version two. The order matters and the destination does not change.
No time figures are claimed as measurements
Twenty minutes for a task, five minutes of capture, an hour from memory — these are working figures describing relative cost, not findings. Nobody was timed. They are in the post because "smaller" and "cheaper" are unactionable and a number a reader can test against their own next task is not.
These documents are starting structures for you to build on, not professional or legal advice. Anything that will have legal or contractual effect, or that varies by jurisdiction, should be reviewed by someone qualified in the relevant place.
Common questions
- How do we write procedures when nobody has time?
- Stop trying to document a process and document one task instead — the narrowest unit somebody could hand to another person and walk away from. It fits in twenty minutes and it is finished the same day, and a finished narrow document beats a stalled broad one on every measure that matters.
- Is it faster to write a procedure from memory or during the work?
- During the work, by a wide margin. Writing from memory takes an hour, comes out in the wrong order, and silently omits the steps the person no longer notices doing. A line per step written as the task happens costs about five minutes on top of work that was happening anyway.
- What about tasks that only happen once a year?
- Write it the next time it happens rather than now. Put a line in the calendar entry saying to capture it, and accept the document does not exist until then. A wait of months is still faster than a reconstruction nobody starts.
- Does the first version have to be properly formatted?
- No, and holding it to that standard is where much of the time cost comes from. A numbered list with no control block is already worth more than the nothing it replaced. The one thing worth doing properly from the start is having somebody else follow it once without help.
- How many procedures should we write at once?
- One, then stop. One document a month that is current and used beats a library written in a single push and stale by spring. Start with the task where one person is the only one who can do it and the work stops if they are away.