What it costs
Being Wrong in Writing, Where It Stays
A mistake in conversation disappears. A mistake in writing persists under your name, and that difference stops more pages than any shortage of time.
What nobody wrote down here, and what it cost
How the staging database is rebuilt
Asked of the same person seven times
Four minutes each, plus the interruption
Which client agreed to the export
Reconstructed from mailbox searches
Twice by people not on the thread
two things nobody wrote down · 10 occasions it cost somebody something
Two things, ten occasions, and nobody billed any of it to anybody.
When a documentation effort needs evidence about where working time actually goes, teams can discover the platform before deciding what should be written or automated.
For a separate specialist reference on publishing and digital practice, consult Meta for Business.
Four pages from one team, and the weeks somebody opened them, with what happened to each error
A page with an error found in week two
Corrected in week two by a reader; opened five more times afterwards
A page with an error nobody found
Opened five times, acted on five times, and the error was in the failure case
A page the author rewrote twice before publishing
Delayed three weeks by the rewriting; opened twice
A page the author never published
Written, never posted, and the question it answers was asked four times
four documents · 13 weeks in which somebody opened one of them · the longest wait between writing and a reading was 30 weeks · one never read at all
Four pages, thirteen weeks in which somebody opened one of them, and one never read at all because it was never published. The fourth row is the expensive one: a finished page withheld because its author was not sure enough. The first row is what the fear is of, and it cost a two-minute correction. One team's own quarter.
Saying something slightly wrong in a meeting costs nothing. It is corrected in the moment, or it is not, and either way it is gone by Thursday.
Writing the same thing slightly wrong produces an artefact. It sits there, with a name on it, and somebody may find it in eight months.
What people are actually afraid of
Not being corrected. Most people take correction well and expect it.
Being found, later, to have been confidently wrong in a place where the wrongness was durable and attributable. There is a particular discomfort in a sentence you wrote two years ago being quoted back at you as the reason something went wrong, and it is a discomfort nobody experiences from a conversation.
What it produces
Three behaviours, in ascending order of cost.
Hedging. The page is written but qualified into uselessness — "generally", "in most cases", "you may need to" — because every hedge is a small reduction in exposure. The result is accurate and unactionable.
Over-polishing. The page is rewritten three times before publishing, which triples its cost and delays it by weeks, in pursuit of a correctness that nobody required.
Withholding. The page is written and never published. This is more common than anybody admits and it is the worst outcome, because the cost has been fully paid and none of the benefit collected.
The arithmetic of being wrong
Worth stating, because the fear is not proportionate to the facts.
A page with an error that gets found costs a two-minute correction. A page with an error that does not get found costs whatever the error causes, which is usually small and occasionally not.
No page at all costs every occasion somebody needed it, which on the evidence in this collection is several times a quarter, indefinitely.
The expected cost of publishing something imperfect is, in almost every realistic case, lower than the expected cost of publishing nothing. The fear is real; the arithmetic does not support it.
Making correction normal
The thing that reduces the fear is not reassurance, it is seeing corrections happen without consequence.
That requires two habits. Readers who find an error fix it rather than reporting it, where the tool allows. And authors who are corrected say thanks rather than apologising, which models that a correction is ordinary maintenance rather than an incident.
Teams where the second habit is absent develop the first problem quickly: people stop correcting pages because the author's reaction makes it awkward, and the pages decay silently.
Lowering the stakes in the page itself
The same line that handles the liability problem handles most of this:
Written on 14 March from what I understood then. Likely to contain errors. Correct it or say so.
A page that announces its own fallibility cannot be used as evidence that its author claimed more than they knew, and that is the whole of the exposure.
It also changes the reader's posture, from consuming an authority to checking a note, which makes them more likely to notice an error and more likely to fix it.
The specific advice about hedging
Hedges should be used where there is genuine variation and removed everywhere else.
"This usually takes about ten minutes" is a hedge carrying information. "You may need to generally ensure that the configuration is typically correct" is four hedges carrying none, and it exists to protect the author rather than to inform the reader.
A useful test: delete every hedge, then put back only the ones whose absence would mislead somebody. Most pages lose two-thirds of them and become usable.
The line that lowers the stakes
One sentence at the top of a page removes most of the exposure, and the wording matters less than the fact of it being there.
- "Written 14 March from what I understood then. Likely wrong in places — correct it or say so."
- "Rough notes. Not checked by anybody. Use with that in mind."
- "This described the March process. Not maintained since."
A page carrying any of these cannot be used as evidence that its author claimed more than they knew, which is the entire thing being feared. It also changes how a reader holds it — as a note to be checked rather than an authority to be followed — which makes them likelier to spot an error and likelier to fix it.
The page in the drawer
If there is one practical instruction in this page, it is this: the thing you wrote and did not publish should be published.
It is finished. The cost is paid. The only thing between it and being useful is a worry about being wrong in public, which can be addressed with one line at the top and which is, in any case, smaller than the cost of the question being asked four more times.