What it costs
What Not Writing Costs, and Who Is Billed
The cost of an absent document is real, large and distributed across people who never connect what they paid to what was missing.
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.
Teams testing a new knowledge workflow can the official information to establish a baseline, then compare the result after ownership and review rules change.
For a separate specialist reference on publishing and digital practice, consult Instapage.
One team's quarter, with every cost of an absent document traced to whoever actually paid it
How the staging database is rebuilt
Asked of one engineer seven times
Paid by that engineer, in interruptions
What the deploy order is
Guessed, then corrected by whoever noticed
Paid by nine different people, twenty minutes each, and once by a customer
Why the import runs twice
Asked at incidents, never answered
Paid by whoever was on call, in longer incidents
Which dashboard is authoritative
Asked in the support channel
Paid by five people, and once by a client who received a wrong figure
Which client agreed to the custom export
Reconstructed from mailbox searches
Paid by two people who were not on the original thread
five things nobody wrote down · 29 occasions it cost somebody something
Five absences, twenty-nine occasions. The costs fall on at least fifteen different people, a customer and a client, in amounts between four minutes and most of a morning. Not one of them appears in any record as a cost of anything. One team's own quarter, traced afterwards.
A page not written has a price, and the price is paid in full by somebody.
It is simply not paid by the person who would have written it, is not paid at one time, and is never recorded as the price of anything, which is why the whole subject feels optional.
Where the costs land
Five places, and the list is worth having in full because people usually think of only the first.
The repeated question, paid by whoever is asked, in interruptions.
The reconstruction, paid by whoever needs the answer and cannot get it, in time spent working it out from the artefact.
The wrong guess, paid by whoever acted on it and by whoever cleaned up, which occasionally includes a customer.
The re-argument, paid by everybody in a meeting where a settled question is reopened because the settlement was never recorded.
The permanent loss, paid by the organisation, when the person who knew leaves and the reason is gone.
Why none of it is billed
Each instance is small enough to pass as ordinary friction.
Twenty minutes working something out is not experienced as a cost of missing documentation. It is experienced as doing your job, and it looks exactly like doing your job, and nobody writes "spent twenty minutes because nothing was written down" in any system.
The absence of a billing mechanism is not an administrative oversight. There is no plausible mechanism: you would have to know, at the moment of paying, that a document would have prevented this, and the whole point is that you do not know what is missing.
The one place it becomes visible
When somebody leaves.
A departure converts a distributed invisible cost into a concentrated visible one. Suddenly there is a list of things nobody knows, and people can name them, and the cost of finding out is estimated and sometimes paid in full.
This is why departures produce short bursts of documentation activity, and why those bursts do not generalise: the cost became visible for two weeks and then went back to being invisible.
Estimating it roughly
An estimate is better than nothing and is easy to make badly, so it is worth doing conservatively.
Take the tally from the previous page — the list of questions asked and how often. Multiply by a modest guess at the cost of each, including the interruption. Do not multiply by a recovery factor, do not extrapolate to a year, and do not convert to money.
A plain statement — "these five absences cost thirty-two interruptions and about nine hours this quarter, spread over fifteen people" — carries more weight than any figure with a currency symbol in front of it, because the first can be checked and the second invites an argument about the multiplier.
Who to show it to
Whoever decides what the team works on.
The purpose is not to prove a point but to move a specific page onto the list of things that get time. The argument lands because it is concrete: this question, this many times, these people.
A general argument about documentation value does not land, because it has been made before and everybody already agreed.
The honest counter-argument
Some of this cost is cheaper than the page would have been.
A question asked twice in a year, by people for whom it took four minutes, cost eight minutes. A page answering it costs forty and needs maintaining. Writing it would have been a loss.
The reason to count rather than assert is precisely this: counting tells you which absences are expensive and which are fine, and a team that writes everything is wasting about as much as a team that writes nothing, with more effort.
Where the costs actually land
The list is worth having in full, because most people think of the first item and stop, and the first item is the smallest.
- The repeated question — paid by whoever is asked, in interruptions.
- The reconstruction — paid by whoever needed the answer and worked it out instead.
- The wrong guess — paid by whoever acted and whoever cleaned up, occasionally by a customer.
- The re-argument — paid by everybody in a meeting where a settled question is reopened.
- The permanent loss — paid by the organisation when the person who knew leaves.
Five rows, five different payers, and not one of them records the payment as the cost of a missing document.
What the counting is really for
Not to prove that documentation matters. Everybody agrees already, as the first section of this collection argued at length.
It is to identify the five or six specific things whose absence is actually costing something, out of the hundreds of things that are also undocumented and are fine. That list is short, it is unexpected, and it is the only documentation backlog worth having.