Two questions get mixed up at the end of every training. Is this person ready to work alone, and did the training work? They are answered at different times, with different evidence, and an area that only asks the first one never finds out whether its method is any good.
The release
The release is answered at the station, at the end of the follow-up step. The trainer observes one cycle, counts the key points that are followed, judges the first part and signs with a date. For a station where the person judges a test or an inspection result, the quality function signs as well.
Three names on one line are two too many. A2 settled on a proposer, a signer and a recorder, wrote down who they are, and put the level in the matrix within a week of the signature.
The effectiveness check
Four weeks after the release the area reads two numbers for that person: the defect rate of their output in those four weeks, and how long it took them to reach the standard cycle time of the station. Both are compared with the area, which runs at 0.9 percent.
A2 has thirteen people to compare, six trained by watching in the year before and seven trained in four steps with a breakdown in the year after.
| Measure | Learned by watching | Four steps |
|---|---|---|
| Weeks until the standard cycle time is reached | 6 | 3 |
| Defect rate in the first four weeks, percent | 3.8 | 1.4 |
| Releases signed by the planned date, of the group | 2 of 6 | 7 of 7 |
Three weeks of output per person and 2.4 percentage points of defects are not a soft benefit. At four new or moved people a year they are EUR 9,688, and they are the part of the business case that keeps paying after the coverage gaps are closed.
The same two figures across everybody trained this way answer the third question, the one about the method itself. Read them every quarter. If they do not move, the breakdowns are thin or the sessions are being cut short, and both are visible in the training log.
Six key figures
| Key figure | How it is counted | Target in A2 |
|---|---|---|
| Coverage | Covered slots divided by required slots | 100 percent |
| Slots below the minimum | Station and shift slots under the agreed minimum | 0 of 16 |
| Trainer coverage | Stations with a person at level 4 | 8 of 8 |
| Plan adherence | Released by the planned date, of those planned | Above 90 percent |
| Days to release | First session to the signature | 30, or 45 at a test station |
| Start-up defect rate | Defects in the first four weeks, of that output | Below 1.5 percent |
Coverage, the slots below the minimum and plan adherence belong next to output and quality in the monthly area review. A key figure that is only reported when somebody asks for it is not managed.
The upkeep routine
A matrix decays faster than any other list in the area. Four habits keep it alive.
Within one week, the owner records every release, transfer, leaver and expired validity. Every month, the area manager reads coverage, gaps, plan adherence and the releases that run out in the next three months. Every quarter, three cells at level 3 are checked against the evidence, one per shift. Twice a year, the rules themselves are reviewed: the levels, the validity, the station list and the trainer list.
A2 went from 78 percent coverage and six slots below the minimum to full coverage in six months, and from 26 to 78 points on the twenty-five statement check. The part that holds is the smallest: no release, no station, and every change in the matrix within a week.