When the Founder Is the Memory System

You are tired of being the only one who remembers why you decided things.

Not the only one who can decide. The only one who remembers the reasoning: why that client is on different terms, why you stopped doing the thing everybody keeps suggesting, what you promised the landlord in 2023, which supplier you will not use again and why.

This is the most common failure in a growing company and it does not look like a failure. It looks like commitment. It gets praised.

Why it holds

For the first few years it is genuinely efficient. One person holds everything, decides fast, and never has to write anything down because they are always in the room.

Then the company grows and the same arrangement becomes the constraint:

Every non-trivial decision routes through one person. Not because they insist. Because they are the only one who has the context to decide well, and everyone knows it.

The reasoning is invisible. People see the decision, not the argument. So they cannot extend it. Faced with a new case that is nearly the same, they have to come back and ask.

Questions arrive faster than they can be answered. The founder becomes a queue. Work stops while people wait, and the waiting is invisible in every report you look at.

Nobody can be handed anything whole. You can delegate the task. You cannot delegate the fifty pieces of context that make the task go right, so it comes back wrong and you conclude that delegation does not work here.

It is unsafe. If that person is ill for a month, the organization does not slow down by a tenth. It stops making decisions.

Measure it before you fix it

Everyone in this situation already knows it is happening. What almost nobody has is the size, and without the size you cannot tell whether an intervention worked.

1. The interruption count. For one normal week, have the founder tally every time somebody asks a question only they can answer. Not tasks. Questions of context or permission. Twenty is a heavy week. Sixty means the organization has no memory outside one head.

2. The queue. For the same week, have everyone else log every time they were blocked waiting on an answer, and for how long. The founder's tally and everyone else's rarely match, and the gap is the point: the founder feels interruptions, everyone else feels waiting, and neither sees the other.

3. The bus test, done honestly. Pick three live commitments: a client agreement, a supplier arrangement, a promise made in a meeting. Ask someone else to find the terms and the reasoning, without asking the founder. Time it. If the answer is "I would have to ask", you have measured the exposure.

4. The re-decision count. How many times in the last quarter did you discuss something you had already settled, because nobody could find what was decided or why? Every one of those is the same meeting held twice.

The first three things to move

Do not start by delegating decisions. Start by making the reasoning findable, because a decision without its reasoning cannot be extended by anyone else.

1. Decisions, with their reasons. Not a task list. A short record of what was decided, by whom, when, and why — including what you decided against. Two paragraphs is enough. See Decision Records.

2. Commitments and their terms. Everything you are contractually or morally on the hook for, with dates and the person accountable. Most organizations do not have this anywhere. It is where the nasty surprises live.

3. The exceptions. Every rule you have quietly broken and the reason. "We normally require a deposit; this client is exempt because X." Exceptions are almost never written down and are the single most common thing only the founder knows.

What actually changes it

Answer once, in writing, where it can be found again. The next question is the same question. If you answer it in a direct message, you have committed to answering it forever.

Give away the decision, not the task. Delegating a task keeps the context with you. Delegating a decision, with a stated boundary, moves the context. See The Advice Process.

Say your reasoning out loud, especially when it is obvious to you. The reasoning is the transferable part. The conclusion alone teaches nothing.

Let someone decide differently and back them. The first time somebody makes a call you would not have made and you support it publicly, the organization learns that decisions really have moved. Until then everyone is waiting to see.

Where it goes wrong

Buying a tool first. A wiki with nothing in it is not memory, and a tool cannot make anyone write down reasoning they have never written down. Start with the three lists above and a habit.

Writing everything. Total documentation is a fantasy and the attempt collapses in about six weeks. Write decisions, commitments and exceptions. Nothing else has this problem.

The founder writing it all themselves. That is one more thing routed through the bottleneck. Whoever asked the question writes the answer down; the founder corrects it.

Treating it as a personal failing. It is not a discipline problem. The arrangement was correct at eight people and stopped being correct at twenty, and nothing marked the transition.

What good looks like

Somebody two years into the company answers a question about why a client is on unusual terms, correctly, without asking anyone.

The founder takes two weeks off and decisions keep being made. Some of them differently. None of them stopped.

Try this week

Run the interruption count for five days. Just the tally, no analysis.

Then take the three questions that came up most and answer them once, in writing, somewhere the next person will look.

That is the whole method. The number is what tells you it is working.