How to document processes (and survive a generational handover)

The knowledge in your head is a liability

It feels like a strength: you just know how the whole business works. Every process, every quirk, every workaround lives in your head, and it runs. That feeling is exactly the risk, because it is invisible until the day it isn’t.

Most founders can name their own Peter in a heartbeat — the one or two people without whom a couple of critical processes simply stop. Everything can be repaired, of course. The question is the bill: in euros, in management time spent firefighting, in customer goodwill, and in the patience of the rest of the team who watched it happen. How the year-end gets prepared, how a big customer likes their delivery notes, where the supplier contract is filed and what is actually in it, the login to the payment system — it all works fine until that person is on holiday, off sick, or hands in their notice. Then the honest answer to “what happens now?” is often “we improvise, badly.”

The finance world calls this key-person risk. The plain version is a single question: what happens to this business on the Monday morning the wrong person doesn’t come in? And it is not only the dramatic version that costs you. Every day, if only one person can do a thing, it always comes back to that person — so they never get to the more valuable work, and the business can never grow past them.

Documentation is what lets the business run without you

The point of writing things down is not tidiness. It is freedom. A documented business can delegate a task and trust it gets done, onboard a new hire in days rather than months, keep running while someone is away, and grow beyond the founder’s personal capacity.

Here is the blunt test. Can you take two weeks off — properly off, phone in a drawer — without the business wobbling? Could you lose your phone on safari in Africa and genuinely be unreachable for ten days? If the answer is no, then, with respect, you do not own a business; you own a job. Documentation is the first step from the second thing to the first.

None of this means turning your SME into a corporate bureaucracy with a binder for everything. It means the handful of things that currently only work because a specific person is in the building work without them too.

Start where it hurts: name them, rank them, schedule them

The instinct is to document the easy, well-understood tasks first, because they are quick to write up. Resist it — that is precisely the material you least need on paper. Work in three moves instead.

Name them. Have each department list every process it runs — just the names, no detail yet. You cannot rank what you have not written down, and the bare list is usually a small revelation in itself. So push for a genuinely complete one, and notice what quietly doesn’t make it: the process someone leaves off — the one that keeps them a little indispensable, the long one nobody wants to write up, the messy one full of quirks — is usually the very one you most need to capture.

Rank them by risk. Put the critical ones in a simple table and, for each, ask three questions: who is the only person who can really do it, how often does it happen, and what breaks if they are gone?

The coffee cupboard does not need a procedure. The year-end, the key account and the payroll run do — because each sits on one person, and each does real damage if that person is unreachable. Document the red rows first and you have removed most of your risk with a fraction of the effort.

Schedule them. Turn the ranking into a timeline — which process gets written up in which week — so the job has dates instead of good intentions. As a rule of thumb, reach first for the single points of failure (one person, no backup), the compliance-critical ones (tax, payroll, contracts), the high-frequency tasks where a mistake is expensive, and the founder-only jobs quietly keeping you from higher-value work.

How to document a process well

Now the writing itself. The reason processes never get documented is that everyone pictures an enormous manual and quietly decides not to start. So do not picture a manual. The unit of work is one process, one checklist, one afternoon — and there is a clear bar to aim for.

The standard is this: another competent professional in the field — or, stricter, at least one other person already on your team — could perform the task correctly from the documentation alone. It is fine if they would need up to three times as long the first time, working only from your notes. Slower is allowed; unable is not. That, not elegance, is the test.

Capture it while you do it. The person who does the task is the person to document it — do it once, slowly, writing each step as you go, or record your screen and write it up afterward. Do not try to reconstruct it later from memory in a meeting room.

Every process worth the name has the same four parts. Get these down and you have a real procedure:

Inputs — and where they come from. What does the task need to start, and where does each piece come from? If getting the inputs is itself work, that is step one of the process. When an input comes from another department, name the department, not the person, and the deadline you agreed for it to arrive: “booking data arrives from Accounting as a structured CSV by the second working day,” not “Maria sends over the Excel at some point.”

The steps — how the input becomes the output. Numbered and concrete, pitched at the level of someone in the same profession. Not “Controlling turns Maria’s file into the report management wants,” but: “enter the order list, add the client name by looking up the purchaser’s email against the customer master data, then show sales per client per week in a pivot.” Any controller could follow that — including one who has never worked here. Not everyone in marketing needs to; the reader of a controlling process is a controller.

Outputs — and where they go. What the task produces and where it lands: saved to this folder, emailed to that department, posted into that system. A process whose output nobody can find is only half done.

The deadline. When it must be finished — stated in working days (“by the fifth working day of the month”), not calendar dates, since the work happens on working days. Reserve hard dates for the genuinely fixed: rent on the 1st, the VAT return on the last day of the month.

One balance to strike: document what is unique to your systems and files in detail — that the customer number is called “Kundennr” because the source system is German, that one report hides a quirk in column F — but assume the reader knows the standard tools of their trade. Any controller knows how to use Excel; you do not need to teach it. No black box, but no re-teaching the basics either.

Finally, give each document a home, an owner and a date. Put it where people actually look — a shared drive or a simple wiki, not someone’s inbox — name who owns keeping it current, and date it, because an undated document rots quietly until it is wrong. And do not wait for perfect: a rough version in the shared folder today is worth more than the perfect one still in your head. You can always improve a document that exists.

Expect resistance — including your own

There is a predictable last obstacle: people do not want to write their processes down. Sometimes it is honest overwhelm; sometimes it is quieter, because being the only one who can do a thing feels like job security. Either way, the two excuses you will hear are the exact opposite of the truth.

“It’s too complicated to write down.” Then it is the most important one to write down. A process so tangled that only one head can hold it is the single greatest risk on your list — “too complicated” is the reason to start, not the reason to stop.

“It only takes me ten minutes — I’ll do it on holiday too.” You deserve an actual holiday. And a task that depends on you being reachable on a beach is not covered, it is exposed — ten minutes for you can be a lost day for everyone else, precisely when you are hardest to reach.

And even wishing everyone the very best of health: who is to say you will not be flat on your back for ten days with a stubborn flu? You document processes for the same reason you insure the building — not because you expect the fire, but because you cannot afford to be wrong about it.

The real test: handover, succession and sale

Everything above matters more on one particular day — the day the business changes hands. To the next generation, to a new managing director, or to a buyer. That is when undocumented knowledge stops being an inconvenience and starts being expensive.

The generational handover. In a family business the founder quietly assumes the successor “just knows” how it all works, because they grew up around it. They don’t. The tacit knowledge — which supplier to call in a pinch, why you never discount for that customer, how the seasonal cash squeeze is handled every year — is exactly what was never written down, and exactly what the next generation needs on day one. Hope is not a handover plan.

The sale. A sale is the unforgiving version. A buyer pays for a business that runs; they discount, heavily, for one that runs on you. Due diligence is essentially a stranger trying to work out how the business actually functions — and every “oh, only Jan knows that” is a mark against the price or a reason the deal stalls. Documented processes are, quite literally, worth money.

The move is the same in every case: document while it is calm, years before you need to. A handover you have prepared for is an orderly transfer of a working machine. One you haven’t is a fire sale of knowledge you can only sell once — and usually at a discount.

What good looks like

You can hold your own business against this list:

• You have named every process, ranked them by risk, and put dates against documenting the critical ones.
• The processes that would hurt most are written down — not just the easy ones.
• Each critical process states its inputs, steps, output and deadline, and names an owner and a date.
• A capable professional new to the task could run it from the notes alone — slower at first, but without asking you.
• The documents live somewhere everyone can find, not in one person’s inbox.
• No single critical task depends entirely on one person being reachable.
• The financial routines — month-end, reporting, payroll, the year-end with your accountant — are documented like everything else.

Frequently asked questions

You could do this yourself. It is the one job that is never urgent — until it is.

None of this is difficult. Name the processes, rank them, write the risky ones down as checklists, give them owners, keep them current. Any capable owner can do it.

The catch is that it is never the most urgent thing on the desk. Daily operations take the top three spots on the to-do list, and “document our processes” sleeps peacefully at number seven — important every single day, urgent on exactly one: the day the person who knew it leaves, retires, or falls ill. So it waits, and waits, until the day it is too late to do calmly. That is the specific reason this obvious, valuable job so rarely gets done.

That is where we come in. You do not want to be figuring out how to set up the project — we have done it before. We can provide the templates, guide the team through it, and run a first round of quality control on the processes you produce. And where it helps, we bring the outside perspective and the gentle, well-meaning pressure that keeps it moving — so it does not turn into a forever-project.

On the financial side we can of course go deeper — helping you document, evaluate and optimize your finance processes: the month-end close, the reporting routine, the booking rules with your accountant, the cash flow process — so those never live in a single head.

Off your plate. A business that runs on process, not on you — worth more the day you decide to step back, and calm enough that you can finally take that holiday.