Every growing business runs on a body of knowledge that no system holds. It is knowing which customer needs their invoice formatted differently, which supplier will deliver early if asked on a Monday, which exception to the standard process is safe to approve without escalating. None of it appears on a balance sheet. All of it walks out of the door when the person holding it leaves.

This is institutional knowledge: the accumulated context that lets a business operate smoothly rather than by the book. Because it accumulates inside individuals rather than inside systems, every departure withdraws capital the business never knew it had. The recruitment fee is recorded. The lost judgement is not.

The argument here is not that businesses should attempt to document everything. It is that knowledge transfer should be treated as an operating process — mapped, scheduled and rehearsed while the knowledge is still in daily use — rather than as an exit formality completed in a final fortnight.

Why knowledge concentrates in a few people

Speed favours the person who already knows the answer. When a question arrives, the fastest route is to ask the colleague who has handled it before. That is efficient in the moment and cumulative in its effect. Each query resolved from memory rather than from a system reinforces the concentration, until a handful of experienced people are effectively functioning as the organisation's memory.

Exceptions accumulate in the same way. A customer receives special treatment because of something that happened years ago. A supplier is trusted to deliver early because a relationship was built patiently. A report is reconciled with a private spreadsheet because the official version has never quite matched. None of this is written down, because writing it down takes longer than answering the question, and the answer is sitting in the room.

Growth and specialisation make the problem sharper. In a small team, everyone overhears everything; context is distributed almost by accident. As the business adds people, functions and customers, the volume of context grows faster than the mechanisms for sharing it. New joiners learn their role from whoever is nearest, and what they learn is that person's version of the job, including its shortcuts.

This is not a failure of diligence. It is what happens when a business optimises for today's answer. But a single point of knowledge is as much a structural fragility as a single point of failure in a system, and it deserves the same attention.

The full cost of a departure

The visible costs of losing someone are easy to name: the recruitment fee, the notice period, the salary of the replacement, perhaps a contractor to cover the gap. These are the lines a finance team can see and a budget can absorb.

The larger costs are the ones that never get coded to the vacancy.

  • Lost decision context. The reasons behind past choices leave with the person. Why a pricing exception exists, why a supplier was dropped, why a process carries an odd step. Without the reason, successors either preserve the step superstitiously or remove it and reintroduce the problem it was designed to solve.
  • Rebuilt relationships. A customer relationship held by one individual is, in practice, an asset of that individual. The new owner starts again, and the customer notices the reset.
  • Repeated mistakes. Problems solved years ago are solved again, at full cost, by people who never knew the first solution existed.
  • Manager time. Weeks of a manager's attention spent reconstructing what left, answering questions that a colleague used to answer across a desk.
  • Slower recovery. The business keeps operating, but in a lower gear for a period nobody has budgeted for.

Individually these are irritations. Together they can set a team back a quarter, and the setback rarely appears anywhere as a single identifiable number.

Documentation is not the same as transfer

The standard response to a resignation is to ask for a handover document. It rarely delivers what the business needs. A document written in the final fortnight captures what a departing person can recall under time pressure, in a form nobody has reviewed and few will maintain. Written procedures that nobody reads give a false sense of safety, which is worse than an acknowledged gap because it looks like coverage.

The knowledge that matters most is largely judgement: which exception is safe to approve, when to escalate, how to read hesitation in a long-standing customer's email. Judgement transfers through use rather than description. It moves through pairing, structured shadowing and supervised overlap, where the successor does the work and the experienced colleague observes, corrects and explains. Because the work is live, the cases that matter actually come up.

Mapping where the business is exposed

Before designing any transfer, it helps to know where the concentration sits. A useful question for each role is this: if this person left on Friday, what would we be unable to do by the end of the following month, and how long would it take to rebuild? Where the answer is uncomfortable, the role is a single point of knowledge.

That exposure belongs on the same risk register as customer concentration or supplier dependency — reviewed periodically, discussed openly, and reduced deliberately. Framing it as a personnel matter tends to postpone the conversation until the resignation letter arrives, which is the worst possible moment to start.

From exit formality to operating process

Transfer works better when it is built into ordinary events rather than reserved for departures. When a project closes, the team records what was decided, which alternatives were rejected and why. When someone is promoted, the transition includes a scheduled overlap rather than a single handover meeting. When a new person joins, shadowing is part of the plan from the first week, not an informal arrangement that depends on who happens to be free.

What to capture is a short list: decision rationale, customer and supplier context, known exceptions, and the judgement calls that never make it into a process document. The aim is not a complete record of everything a person knows, which is neither possible nor useful. It is to capture what would be expensive to rebuild.

Knowledge transfer is cheapest when it is routine and most expensive when it is urgent. A business that maps its exposure, builds overlap into normal role changes and treats transfer as a process rather than a formality keeps far more of what it knows. The knowledge still lives in people, which is where judgement belongs. The difference is that it lives in more than one of them.

← PreviousBack to Articles