The work only one person can do is the work most at risk.
Picture the week a business runs while one person is away. Not a quiet week, a normal one, with the usual load of work moving through. If that week goes fine, the business has processes. If it wobbles, if things sit half-done and clients wait and everyone braces for the return, then what the business has is a person doing a set of tasks well. Those are not the same thing, and the gap between them is where most of the operational risk in a small business genuinely lives.
We call it the holiday test, and it is more revealing than any org chart. A process is something you can hand over. If you cannot hand it over, it was never a process; it was one person's competence, held in their head, dependent on their presence. That competence is real and valuable. It is also fragile, and it does not scale, and it cannot be sold with the business.
Why the best people hide the problem
The uncomfortable part is that capable people mask this risk rather than expose it. A strong operator absorbs the gaps. They remember the exception to the invoicing rule, they know which supplier to call when the usual one is slow, they carry the sequence of steps that never got written down because they never needed writing down. Everything runs smoothly, so nothing looks broken.
Then they take leave, or they resign, or they simply have a bad fortnight, and the smoothness turns out to have been them. The business did not have a system for that work. It had a person, and the person was the system. The better they were, the less anyone noticed how much was riding on them.
What a process actually is
A process is the work made independent of the worker. It is repeatable by someone who was not there when it was invented. The test is not whether it is written in a specific format or stored in a specific tool. The test is whether a competent person, new to the task, could pick it up and get it right without the original owner in the room.
That reframing matters, because most people hear "process" and picture bureaucracy: sign-offs, forms, a manual no one reads. Real process is lighter than that. It is the difference between "Sarah handles the month-end close" and "here is how month-end close is done, and Sarah happens to do it this month."
Documentation is one page, not ninety
This is where good intentions usually stall. Someone decides the business needs documentation, pictures an enormous manual covering every scenario, feels the weight of it, and never starts. The ninety-page manual is a fantasy anyway. Nobody writes it, and if they did, nobody would read it, and it would be out of date before it was finished.
What actually works is smaller and far more useful: the one clear page that lets the next person get it right. Not every edge case, the main path. The inputs, the steps in order, the decisions and how to make them, the common thing that goes wrong and what to do about it, and who to ask when the page runs out. A page like that can be written in an afternoon. It captures the eighty percent that is done the same way every time, and it turns "only Sarah knows this" into "anyone can follow this."
Consider a firm where client onboarding lived entirely with one team member. Every new client went through her: the welcome sequence, the file setup, the systems access, the first invoice. It worked beautifully, right up until she went on leave during a busy intake month and three new clients stalled in silence. Nobody else knew the order of operations. The fix afterward was not a manual. It was a single page: the seven steps, in order, with a note on the two that trip people up. The next time she took leave, onboarding kept moving, and she came back to a business that had run without her instead of a backlog that had waited for her.
The best time to write it down
Here is the pattern we see again and again. The page gets written in a crisis, when someone has already left or is already away and the knowledge is walking out the door. That is the worst possible time. The person who holds it is gone or checked out, the pressure is high, and whatever gets captured is rushed and incomplete.
The best time to write it down is before you need it. When the person who does the work is present, calm, and doing it well, they can explain it clearly and check the page against reality. Writing it down then costs an hour. Writing it down in a crisis costs a stalled week and a client relationship, and you still end up with a worse page.
There is a subtler benefit too. Writing the process down often improves it. When someone has to articulate the steps for another person, they notice the ones that only make sense because of history, the workaround that could be removed, the approval that no longer serves a purpose. The act of making work handoverable tends to make it better.
The test worth applying
Pick the three things in the business that would hurt most if the person who does them vanished tomorrow. Not the glamorous work, the load-bearing work: the billing, the client handovers, the reporting, the thing that keeps the lights on. For each one, ask whether it is a process or a person. Could you hand it over this week with one page and a short conversation, or does it live entirely in someone's head?
The businesses that hold together when a key person steps away are not the ones with the thickest manuals. They are the ones that wrote the important pages while there was still time to write them well. The work only one person can do is the work most at risk, and the fix is almost always cheaper and simpler than the risk it removes.