Knowledge transfer before retirement: how to preserve 30 years of expertise

There comes a moment when things become uncomfortable. Not on the shop floor, but in the management room. The moment someone says out loud what everyone already knows: in two years, the most experienced operator will retire. And nobody can do what he can do.

That is not an HR issue. That is a business risk.

What actually leaves the business

When a senior employee retires, more than a person leaves. A full knowledge system leaves with them. Thirty years of machine quirks, customer-specific adjustments, exceptions to procedures, and solutions to problems that never appear in manuals.

This type of knowledge has a name: tribal knowledge. It is knowledge that is not documented, not stored in systems, and does not fit inside a two-week training programme. It is why a specific machine is set slightly differently once temperature rises above twenty degrees. Or why one customer always expects a non-standard tolerance that exists nowhere in written notes.

Research by TNO from April 2026 confirms the urgency. In their report Without robotisation, Dutch manufacturing will disappear, they describe how ageing and labour shortages are putting manufacturing under pressure. Productivity needs to rise by fifty percent over the next decade. That target is out of reach if the knowledge that keeps production running disappears with every retirement.

The bill

Let us make this concrete. When an experienced operator leaves without proper transfer, this is what it costs:

The total impact rises to €150,000 to €250,000 per departing expert. If two or three seniors leave within three years, you are looking at half a million euros in risk. Not in line items on an invoice, but in lost productivity, declining quality, and customer impact.

Why it keeps failing

Most manufacturing companies have already tried to solve this once. Usually through one of these three approaches.

• The wiki nobody fills in. It starts with templates and deadlines, then fades after a short period. Not because of resistance, but because operators are technicians working at machines, not behind a keyboard.
• The videos nobody watches. Critical knowledge is buried in long recordings that are hard to search when someone needs one specific answer quickly.
• The consultant who was temporary. Documentation initiatives often produce a report that is outdated within months, while the knowledge fades again.

The pattern is familiar. Not because companies approach it incorrectly, but because the method does not match the people. An operator who has worked with their hands for thirty years will not suddenly switch to writing down expertise in documents.

What sits underneath

The core issue in knowledge transfer is that most of the relevant knowledge is implicit. People have it, but cannot easily write it down. Not because they do not want to, but because it is embedded in experience, intuition, and thousands of repetitions.

An operator can hear when a machine sounds different. They can feel when a product is not right. They know customer X expects a specific deviation in a specific order type. Ask them to write that down and they will say: I just know.

That is not an excuse. That is the nature of implicit knowledge. It cannot simply be converted into documents or instructions. That is why onboarding in manufacturing often takes six to twelve months, not two weeks.

What SMEs experience

The study Production automation in SME manufacturing companies by Midpoint Brabant, REWIN, and Fontys University of Applied Sciences (April 2026) shows that tight staffing is one of the top three challenges manufacturing companies are trying to solve. At the same time, 63 percent of respondents report they are still in the early stages of data and AI maturity.

That combination says enough. The urgency exists. Labour becomes scarcer. But the tools for preserving knowledge often do not fit the shop floor. In small, high-mix environments, tribal knowledge is the difference between smooth production and costly downtime.

Where the shift happens

A shift is happening in how companies approach this issue. The old model was: document knowledge and store it in a system. The new model is: capture knowledge at the moment it appears, in the medium operators already use.

In the same line, TNO highlights in Without robotisation, Dutch manufacturing will disappear that combining digital work instructions with human expertise is a key success factor. Operators do not need to be replaced. Their knowledge needs a place outside their heads.

How to start tomorrow

Knowledge transfer does not have to be a major project. It starts with three steps.

• Map your knowledge holders. In every plant, there are one to three people everyone depends on. If they are not available, problems start. Begin there.
• Keep scope small and concrete. Do not try to document thirty years at once. Start with one machine, process, or customer where dependency is highest.
• Use a method that fits operators. If the method requires typing, logins, or forms, it will fail. The approach needs to be as low-friction as everyday shop floor communication.

The real risk

The risk of doing nothing is not that things fail tomorrow. The risk is failure in two years, when the senior operator has their last day and nobody knows exactly how to set a critical machine under off-nominal conditions.

This is not a dramatic event you can spot in one day. It is slow loss you notice too late: returned orders, machine downtime, and errors by new staff that the senior never made.

Knowledge transfer is not an innovation project. It is risk management. The best moment to start is before urgency forces you to.