Making jar 400 match jar 1

Consistency is the product. A customer who buys twice is buying the promise that the second jar is the first one again, and everything about scale works against that.

What actually drifts

Four things, in roughly this order of likelihood. The incoming material, because crop years and origins differ and a supplier may substitute without telling you. The grind, when a mill is replaced or a screen wears. The blend, through small substitutions made under time pressure. And the fill weight, when a scoop replaces a scale.

None of them announce themselves. All of them are visible in records you either kept or did not.

The reference sample is the control

Keep a sealed sample of every batch, cold and dark, labeled with the batch code. Taste new against old, blind, before it goes into jars.

Without a reference you are comparing against memory, and memory converges on whatever you have been tasting most recently. That is precisely how a line moves a long way in small steps with nobody noticing, including you.

Where the variation enters

Mostly at the boundaries. A new supplier lot. A new batch size. A different person filling. A different room temperature and humidity.

Which means the controls that work are boundary controls: check incoming material against the last lot, re-qualify a blend when the batch size changes, and write down the conditions so that an unusual batch is traceable to an unusual day.

Holding it as you grow

Specifications rather than habits. Grind screen, mixing time, fill target, supplier, and formula version all written down as numbers rather than carried in somebody's hands.

The test is the same as for documentation generally: if somebody competent but unfamiliar followed your written process, would a customer notice. Until the answer is no, growth will cost you consistency.

References