You know the person. They can hear a bearing going before the vibration shows up. They know the three-step restart that isn't in any manual, which setting the setup sheet gets wrong on the blue parts, and why the press runs better on nights if you wait ten minutes after start-up. When they're on vacation, the line is slower. When they retire, it can take months to get back to where you were.
That's tribal knowledge: know-how that lives in people rather than in the plant's systems. Every plant runs on some of it. The trouble starts when nobody knows how much of it sits with one person until that person is gone.
Where tribal knowledge hides
- Restart and recovery tricks. The sequence that clears a jam without a maintenance call, or the "tap it here" fix for a sensor.
- Setup adjustments. The offsets and settings that differ from the setup sheet for certain materials or products.
- What the reason codes really mean. Which downtime reason an experienced operator picks for which situation, and which ones they know are wrong.
- What "good" looks like. The judgement at visual inspection that separates a cosmetic mark from a reject.
- Machine history. "It always acts up after a changeover." "We replaced that cylinder twice last year."
Step 1: Find where it's concentrated
Start with a skills matrix, but read it by position, not by person. For each station, count how many people are qualified to run it on their own. A common cross-training rule is at least three per position, so one absence and one vacation don't stop the station.
The stations that fall short are almost always the hardest ones to learn: final inspection, a test stand, a complicated setup. Those are exactly the ones where knowledge is concentrated. Our attendance and skills page walks through an example line where the headcount report says fully staffed and the coverage view shows a single point of failure.
Step 2: Write it down where the work happens
A binder in the supervisor's office doesn't transfer knowledge. Instructions do when they show up at the station, at the moment they're needed. Put the critical steps, the photos of what good and bad look like, and the non-obvious settings into the work instructions and check steps the operator sees during the job.
Ask the expert to review them, not write them from scratch. Experienced operators are usually better at correcting a draft than producing one, and they'll add the details a new writer wouldn't know to ask about.
Step 3: Turn repeat fixes into standard fixes
When the same person keeps fixing the same problem, the fix is tribal knowledge. Record what was done on the work order or issue every time, so the answer lives in the history as well as in one person's head. If the same failure keeps coming back, open one issue for the root cause rather than closing the same repair again and again.
Step 4: Make the shift handover carry it
A lot of know-how passes between shifts in a two-minute conversation at the machine, and some of it doesn't survive shift change. Notes tied to the machine or product, shown to the incoming operator when they log in, keep the context. They also build a searchable record of what each machine has been doing.
Step 5: Let the data hold the pattern
"It always acts up after a changeover" is a pattern someone noticed over years. When stops, reasons and changeovers are captured automatically, that pattern shows up in the data for anyone to see, including the person who started last month.
None of this makes experienced people less important. The plant keeps what they know, and they get to spend their time teaching instead of firefighting.
A simple pairing plan
For each station that's short of cover, pair the expert with two trainees and give it protected time, a few hours a week rather than "when it's quiet". Have the trainees run the station while the expert watches, then the other way round. Measure progress by coverage: the station is done when three people can run it without help, and the instructions reflect what they learned.
How 10in6 helps
10in6 keeps shop-floor knowledge in the system rather than in people's heads. The skills matrix updates from who actually runs each position, so coverage gaps are visible without anyone maintaining a spreadsheet. Work instructions and check steps show at the Operator Console during the job. Shift notes are tied to the machine and shown at login, and every repair and issue keeps its history.