Short answer
Move what lives in your head into three places at once: named people who can do it, systems that hold it, and documents that get used. Work on the few things that would really hurt to lose, start with whatever takes longest, have your own people write it down, and then prove it with a real absence.
Why don’t the usual fixes work?
Most owners have tried at least one of them:
- The binder. Procedures written in a burst of effort, often by the owner or an outsider, filed and never opened. Knowledge moved only into documents becomes a binder nobody reads.
- The new manager. Everything the owner knows goes into one person. The dependency moves one desk over, and the business is exposed again the day that person leaves.
- The new software. A process nobody fully understands gets automated, and the questions still go to the owner.
Each holds part of the answer. On its own, each fails.
What actually needs to move?
Not everything. Most businesses have a few dozen pieces of knowledge whose loss would really hurt and hundreds that wouldn’t. Rank them by how bad it would be to lose them, how often they are used, and how hard they would be to relearn. Then work only on the top of the list. Working through the trivial items is how these efforts stall.
How do you get knowledge out of an owner’s head?
Asking the owner to write it down rarely works, because the most important parts are things he does without thinking. He can describe the routine; the judgement calls, the exceptions and the unwritten rules are what get left out, and they are usually what matters. Drawing those out takes structured methods used in a deliberate order, and it is a large part of what an operational succession program does.
Who should write the procedures?
The people who do the work. A procedure written by the person who will use it gets used; one written by an outsider gets ignored. And no procedure is finished until someone other than its author has used it successfully.
What order should things happen in?
By lead time, longest first. A second person qualifying for a trade licence, customers getting used to a new contact, a lender getting comfortable with someone other than the owner: these run on other people’s timetables and can take a year or more. They start first, even if something else looks more urgent.
What should the targets look like?
Specific and testable. “Improve documentation” can never be finished. “By month nine, three named people can each price a non-standard job unsupervised, and have done so” can.
How do you know it worked?
You leave. Not for a long weekend with the phone on, but for three weeks, unreachable, with someone logging everything that happens. Build up to it with shorter absences first. Expect the first long one to expose something; that is useful, because it shows exactly what to fix next. A signed log of a successful absence is the Three-Week Test, and it is the strongest evidence a business can offer a buyer or lender.
How long does it take?
Usually nine to eighteen months for the main work. Licences and some customer relationships can take longer. More on timing.
Key points
- Documents alone, one manager alone or software alone each move the dependency rather than remove it.
- Focus on the few critical items, not everything.
- The people who do the work should write it down, and someone else should test it.
- Sequence by lead time, set testable targets, and prove it with a real absence.
Related questions
- My business depends on me too much. What can I do?What owner dependency is, the six places it hides, and the order to fix it in.
- How long does it take to prepare a business for sale?A buyer finds it in about ninety days. Fixing it typically takes nine to twenty-four months.
- What happens to your employees when you sell your business?The question most owners rank above price, and what actually protects your people.