Where stock is going, which measurements are drifting, and what is outstanding.
Shadow mode. These are recommendations nobody acts on. §14 Phase 2: the system proposes, planners work as normal, and both decisions are recorded — so the two can be compared before anything is allowed to change a bin. There is no write path to any ERP, by design (§3.2).
§4.4: correct sizing is bigger sizing. Fixing the lead time does not save money — it spends money and buys availability with it. A screen showing only the availability gain is selling, so the money is on the same row.
§13.5 calls falling adoption the most important alarm in the system: it usually leads the accuracy metrics by weeks. In shadow mode these are expected to be near zero — nobody is meant to act yet.
Read the rules in English → — §9.2 picked a model a planner can argue with over one that scores better, and a policy nobody can read is not one.
§9.5: the stored number never updates by itself; the measurement does. The gap is the signal.
| sized on | loops | measured | part master says | gap |
|---|---|---|---|---|
| release to receipt | 234 | 44.3d | 42.3d | +2.0d |
| part master | 25 | 76.6d | 76.6d | +0.0d |
| supplier notice | 2 | 185.5d | 88.5d | +97.0d |
| consumption reconciliation | parts |
|---|---|
| agrees | 203 |
| posted short | 29 |
| never posted | 24 |
| identity review | pairs | records behind them |
|---|---|---|
| customer | 83 | 125 |
| part | 28 | 56 |
| supplier | 0 | 0 |
§4.3: a match rate and a queue depth always travel together. Either number alone is a misleading claim.