You do not need to document everything. You need a yardstick.
"Document everything" is one of the most repeated pieces of management advice and one of the least followed. Just as well — whoever tries produces a manual nobody reads, and gives up in the third month believing the problem was discipline.
The problem was not discipline. It was having no yardstick for deciding what was worth writing down.
What the field shows when nothing is written
A field study of three mid-sized family businesses found something worth recording: repositories and intranets were not mentioned even once. The operation ran on meetings, conversations and the telephone, and process changes "stayed only in the minds of the employees who took part".
It is by Faoro, Biazutti and Silva, published in 2018 under the title Compartilhamento do conhecimento intraorganizacional e interorganizacional.
A manager summed up the cost better than any diagnosis: "there was a lot of corridor conversation, it was far too informal, and what was said ended up being taken as unsaid."
None of those companies was badly run. They had simply grown the way companies grow — solving this week's problem with whoever was nearby.
The yardstick
You do not document everything, and it is not out of laziness: part of what people know does not fit into text. Michael Polanyi said as much in The Tacit Dimension, in 1966: we know more than we can tell. Whoever tries to write down a person's entire know-how produces a long, useless document.
The yardstick is different, and it fits in one question: you write down what is a risk to leave unspoken. What, if the person disappears, stops the company.
That changes the size of the task. It is not documenting the operation — it is making a short list of what the company does not survive without, and writing only that. In most small companies, it fits into one person's week.
And it is worth saying what this yardstick clears out of the way: knowledge that only lives in somebody's head is not a professional virtue. It is a company risk. The sentence stings because it sounds like it devalues the person who knows. It is the opposite — the person who knows gets to take a holiday.
The test is not that the document exists
Here is the mistake that has people writing plenty and solving nothing: treating the document as the deliverable.
For each critical thing, what you write is "if it stops, how do you restart it". And the test is not that the document exists, nor that it looks tidy. It is this: another competent person carries it out without phoning anybody.
As long as somebody has to call the author, the knowledge is still in the author's head — now with a file nearby, which gives the comforting sense that the problem was solved.
It circulates either way
There is one observation that takes the weight off the decision: knowledge is going to circulate with or without management. "Whether or not the organisation wants to manage it, it will happen" — that is Davenport and Prusak, in Conhecimento empresarial, from 1998.
So the choice was never between recording and not recording. It is between it circulating in the corridor — informal, lost, available only to whoever was in the conversation — or somewhere it stays, and where the next person finds it.
The half almost everybody forgets
There is a second axis, and it almost never enters the account.
It is not only the knowledge inside the company. There is what lives between it and others: what the supplier knows about your process, what the partner agreed to on the phone, what was learned during a visit. Toyota turned exactly that — its supplier network — into its largest knowledge asset, and Dyer and Nobeoka documented how, in a study from 2000 about that network.
In an ordinary company, this knowledge is folklore: it exists, it is valuable, it decides price and deadline, and it is written down nowhere. It disappears when the relationship changes hands — on their side or on yours.
Recording it takes no system: it takes the supplier visit becoming a written paragraph, and the phone agreement becoming an e-mail confirming it.
The test
Think of the hardest person to replace in your company, and answer:
If they disappeared for a month, what exactly would stop — and is there another person who could take over using only what is written down?
Almost every owner can answer the first half. It is the second that tends to go quiet.