A simple example
A pallet was physically moved Tuesday but the computer move was entered Wednesday from memory. Confirming at the destination attacks the first wrong step.
01 · Check the complaint
Signs to confirm on the floor
- Pick shorts despite recorded stock
- Repeated cycle-count adjustments
- Reserve loads cannot be found
- Returns or problems reappear as bad stock
02 · Possible causes
Check more than one possible cause
The first explanation may sound right and still be wrong. Use the measurements in the next section to separate these possibilities.
1
Moves confirmed late or at the wrong point
2
Uncontrolled overflow and problem locations
3
Duplicate/missing integration events
4
Unit-of-measure errors
5
Adjustment closes variance but not cause
03 · What to measure
Numbers that can change the decision
Accuracy by item/location/status dimensionMeasure whether item, quantity, location, and stock status are each correct—not only whether the total units match.
Variance cause and computer-record historyFor each difference, review the moves and counts since the last known-correct record and name the first wrong event.
Repeat variance by product/locationCount items or locations that show another difference after a prior adjustment.
Move confirmation delayMeasure time between the physical move and the computer confirmation.
Open problem inventory ageCount problem inventory still unresolved and record how many days the oldest has been open.
04 · Possible fixes
Choose a fix only after the facts point to it
05 · This week
A practical seven-day starting plan
- 01
Select 20 recent variances
- 02
Reconstruct the last known correct state
- 03
Classify the first wrong event
- 04
Correct one process cause—not only the record
- 05
Trigger follow-up counts on the same risk class
Turn what you measured into a small test
The diagnostic compares practical options and shows which limits could change the answer.