The situation
A poultry processor was losing coating from a frozen breaded product. The loss was not at the fryer, where it would have been noticed immediately, but somewhere between the plant and the plate. Packs arrived with loose crumb in the bottom and the reheated product had bare patches.
Why it is hard
A coated product is two materials with opposing requirements. The core must hold moisture; the coating must lose it. They are joined by a hydrated bridge: the pre dust absorbs surface moisture and roughens the surface, the batter keys into that, and the crumb keys into the batter.
Moisture does not stop moving when the line stops.
It migrates outward from the core through frozen storage, saturates the layer that was holding the coating on, and the bond fails long after anyone was watching. Which is why adhesion is a moisture problem before it is a glue problem, and why thickening the batter makes it worse rather than better: the coating gets heavier without being held on any more firmly.
What was done
The pre dust layer was treated as the control point rather than the crumb. Surface moisture before coating was measured and managed, and the pre dust was specified to take up that moisture and hold it rather than simply to sit on it. The freezing profile was examined, because the rate of freezing decides crystal size and therefore how much structural damage releases water later.
The crumb was changed last, and only for appearance rather than for adhesion.
What it shows
That the component blamed for a coating failure is almost never the one that caused it.
