What is tribal knowledge in manufacturing, and how do you capture it?

Every plant runs partly on knowledge that is not written down anywhere. The setter who knows the old press needs a half-millimetre shim under the fixture before the first part comes out right. The operator who hears a change in the feeder's sound a minute before it jams. The night-shift workaround for a sensor that sticks when the hall is cold. None of this is in the work instruction. All of it keeps the line running.

That is tribal knowledge: what experienced people know about doing the job that the documentation does not capture. It is valuable, it is fragile, and it usually leaves the building when the person does.

This post covers what tribal knowledge is, why it is a risk, how to find who holds it, how to capture it, and how to stop it going stale once it is captured.

What tribal knowledge is

Tribal knowledge is practical know-how held by a small group of people and passed on by word of mouth, if at all. In manufacturing it tends to fall into a few types:

• Setup tricks. The order to tighten clamps so a fixture does not pull, the offset that works on one machine but not its twin, the coil from one supplier that needs a slower feed.
• Early warning signs. The sound, smell, vibration or colour that tells an experienced person something is about to go wrong.
• Workarounds. The things people do because the official method does not quite work: a sensor that has to be tapped, a part that has to be pre-bent by hand, a programme that has to be restarted twice.
• Judgement calls. When to stop the line, when a borderline part is acceptable, who to call about which problem.
• History. Why a step is done a certain way, what was tried before and why it failed.

Some of this should become standard work. Some of it is a symptom of a problem that should be fixed rather than written down. Capturing it is how you find out which is which.

Why it is a risk

Tribal knowledge is fine while the people who hold it are there. The problem is what happens when they are not.

• Retirements. The setter with thirty years on the same machines retires. Their replacement learns by trial and error, and output and quality dip for months.
• Turnover. Agency workers and short-tenure staff take what they learned with them. If the know-how was only ever passed on by shadowing, every new starter begins from a weaker copy.
• Absence. A key person goes on holiday or falls ill, and a station slows down or stops because nobody else knows the trick.
• Line transfers. When a line moves to another plant or country, the documents move with it. The tribal knowledge stays behind with the people. This is where many transfers struggle. See relocation of production lines for more. Coloplast used the same REWO instructions to support a production-line transfer to Costa Rica.
• Hidden problems. A workaround that only one shift knows about can hide a real equipment or design fault. Engineering cannot fix what they do not know exists.

How to find who holds it

You rarely need a formal knowledge audit. A few practical questions usually point to the right people quickly:

• Who gets called? Ask shift leaders and operators: "When this machine stops and nobody can get it going, who do you call?" The same few names come up again and again.
• What slows down when someone is away? Look at which stations lose output or quality when a particular person is on holiday.
• Where does your skills matrix show one name? A station with only one person who can handle problems is a single point of failure. A skills matrix makes these visible.
• What is different between shifts? If one shift has fewer stoppages or less scrap on the same line, they are probably doing something the others are not.
• What is not in the instruction? Ask experienced operators to walk through the current instruction and tell you what is missing. They usually know straight away.

Then prioritise. Start with knowledge that is held by one person, affects safety, quality or uptime, and is at risk of leaving soon.

How to capture it

Asking experts to write down what they know rarely works. They are busy, and much of what they know they do without thinking. They will leave out the very details that matter most. Watching and recording them doing the job works far better.

1. Pick one task. Not "everything the senior setter knows", but one setup, one changeover, one recovery from a jam.
2. Record the expert doing the job. Use a phone. Film the real task at the real station. Ask them to say what they are doing and why as they go, especially at the points where they check, listen or adjust. Short clips per step are easier to use later than one long recording. Our guide to video recording on the shop floor covers framing, sound and lighting.
3. Ask the "what if" questions. After the recording, ask: "What do you do if it does not seat? If the material is from the other supplier? If you hear that noise?" These answers are often the most valuable part.
4. Structure it into steps. Turn the recording into a step-by-step instruction: one clip or photo per step, a short line of text, key values written out, and the reason behind any non-obvious step.
5. Have a second person follow it. Give the draft to someone who has not done the task and watch them follow it without help. Wherever they hesitate or get it wrong, the instruction is missing something. This test catches more gaps than any review meeting.
6. Review it. Go through the result with the expert and with process engineering or quality. Decide which tricks become the standard method, which workarounds should go to engineering as problems to fix, and which steps need a safety or quality check.

Do this while the expert is still there to answer questions. Capturing knowledge in someone's last two weeks is far harder than doing it steadily over a year.

How to keep it alive

Captured knowledge goes stale like any other document. Machines get modified, materials change, and the trick that worked last year may not be needed any more. To keep it useful:

• Give every instruction an owner. Usually the process engineer or the shift leader for that area. The owner is the person who updates it when something changes.
• Let operators flag problems. The people using the instruction every day are the first to notice when it is wrong. Give them a quick way to say so at the station, and make sure someone acts on it.
• Review on a schedule. Check high-risk instructions at a set interval, and always after a process change, a quality incident or a new machine.
• Keep the expert involved. Ask the original expert to check updates while they are still with you. After they leave, the people they trained take on that role.
• Know who has seen what. When a step changes, you need to know which operators have seen the new version and who is still working from memory of the old one.

Where REWO fits

REWO Guides turns recordings of experts into step-by-step instructions from video, images and text, with two-step approval and version history. Operators can flag a problem in an instruction from the station, and Analytics shows who viewed or confirmed which version. For a plant manager, that means the line depends less on a few people. See REWO for plant managers.

Read next: Relocation of production lines · Skills matrix for manufacturing · REWO for plant managers