Making batch 40 match batch 1
Updated
Easier to control than most crafts in this estate, because everything is measurable, and it only works if you actually measure.
What drifts
Raw material concentration. Materials are sold as solutions at nominal strengths, and a lot at the other end of tolerance changes your finished percentage. Which matters twice: for performance, and because regulatory thresholds are stated as percentages of the finished product.
Water, if it is not deionized.
pH, from both of the above.
Viscosity, which follows salt, surfactant lot and shear during mixing.
Fragrance, from lot variation and from addition temperature.
The specification is the control
A target and a tolerance for pH, viscosity, appearance, odor and batch weight.
Written down, checked on every batch before filling, and treated as a gate rather than a note.
A batch outside specification is information, not an inconvenience: it is usually a raw material lot telling you something.
Retained samples
Two bottles from every batch, labeled with the batch number and the date, kept where you can see them.
They are how you answer a complaint about a batch a year later, how you compare a new batch to an approved one, and how you find out that a formula drifts on the shelf.
They cost two bottles and they are the cheapest quality system available.
Checking a batch
pH, viscosity, appearance, odor and total weight before filling.
Fill weight on ten bottles through the run, looking at the spread rather than the average.
Torque on a sample of caps.
And a comparison against the retained sample from the approved batch, side by side, in the same light.