What drives what, and by how muchManufacturing and pharmaUse case, not a client result
Why does yield vary between shifts, lines and batches?
The decision it informs
Which process parameters to tighten, and which line or shift practice to copy.
A dashboard showsLine 3 yields less than line 1.
An investigation tells youWhich settings, materials and handovers explain the gap, and by how much.
Raw material lot
Setpoint drift on key parameters
Shift handover quality
Tool or die age
Ambient conditions
Changeover duration
01
Where the variation livesThe share of the spread explained by each source, so effort goes where it pays.
02
A parameter windowThe range of each key setting that goes with good yield, with evidence.
03
A practice comparisonHow the best shift runs a changeover, written down so it can be copied.
- What if the key setting is held inside its window?
- What if every shift followed the best changeover routine?
- What if the weakest material lot is excluded?
Method 1Variance decompositionHow much of the spread comes from line, shift, material and settings.
Method 2Statistical process controlWhich variation is normal and which is a signal worth chasing.
Method 3Gemba walks and interviewsWhat the operators know that no sensor records, tested against the data.
Data we would look at
Batch and run recordsProcess parameters from PLCsRaw material lots and supplier dataShift rosters and handover notesQuality results and rework
First week: the questions we ask
What counts as a good unit, and who decides?Which settings can the operator change, and how often do they?Which differences between shifts are known, and which are suspected?What would it be worth to close half the gap?