A signature under a training record proves attendance. It does not prove that the instruction works. That is decided at the station, two weeks later.
Train before release
The new sheet hangs at the station only when everyone who does the task has been trained on it. The trainer shows the job with the sheet in view. The operator does the job and names the key points and their reasons. The record takes one line: person, instruction, version, date, trainer.
For the pilot instruction at station A3 of the example plant, nine operators were trained within five working days. The training matrix of the area lists 96 trainings for the 14 class A instructions. After twelve weeks, 91 were confirmed and five were scheduled.
The question here is narrow: does this version of this sheet work?
Count the key points
Two to four weeks after release, the team leader watches one cycle per person and counts the key points that are followed. The check counts what people do, not what they can recite.
| Measure | Before | Two weeks after release |
|---|---|---|
| Key points observed | 108 | 108 |
| Key points followed | 71 | 103 |
| Rate | 66 percent | 95 percent |
The 108 are nine operators times twelve key points. Three key points explain most of the gap at the start: the fitting cone was used by three of nine operators, the tie rods were tightened in the right order by four, and the grease was spread around the full circumference by five. After the release, cone and tightening order were at nine of nine and the grease at eight of nine.
Read the result
- One person misses several key points. This is a training gap. Train again and check again after one week.
- Three or more people miss the same key point. The sheet or the workplace is the problem. Change the wording, the photo, the tool or the layout.
- A safety key point is missed. There is no tolerance. Stop, clarify at once and check the whole team within the week.
- The rate is at 90 percent or above. The instruction works. The next check comes with the regular review.
The second case keeps the plant from retraining people for a fault that sits in the document.
The result figure
The count of key points is a leading measure. The result figure of the task confirms it. Behind station A3 stands the leak test. Its failure rate fell from 4.2 to 1.1 percent within twelve weeks, from 252 to 66 reworks per month. A new operator now works alone after two shifts, down from five.
The whole program of the area, 14 class A and 22 short instructions, takes 117 hours and costs about EUR 7,300 including holders and printing. Less rework at A3 and shorter training bring about EUR 24,400 per year. The payback is 3.6 months.
Keep it current
An instruction starts to age on the day of its release. Six events trigger a change.
- A defect or a complaint that points to a step.
- An accident or a near miss.
- An engineering change: new material, new value, new part number.
- A new tool or machine.
- A proposal from the station.
- The review date: 12 months for class A, 24 months for short instructions.
At the review, the owner watches one cycle with the sheet in the hand and asks three questions. Does the job still match the sheet? Do the values still match their source? Did defects point to a missing key point? A review that changes nothing still gets a date and a name in the register.
Outlook: the sheet on a screen
A screen at the station can show the right variant for the order and a short video of a hand movement. It adds no content. A poor sheet on a screen is still a poor sheet, and scanned text stays unread.
The sensible order is to make the paper system work first. Steps, key points, reasons, pictures, versions and training records then carry over as they are, and the screen only changes how they are shown.