Nobody changed the recipe and the kibble started breaking

A formula that has run clean for two years starts throwing fines.

Nothing on the ingredient sheet changed. Same suppliers, same grades, same certificates of analysis, all in spec. The line settings are where they have always been. And yet the piece coming off the die is more fragile than it was last season, and it is showing up as breakage at the customer's end rather than at yours.

The first place I would look is amylose.

Starch is two polymers. Amylose is linear, and it sets a firm gel with a strong film, which is what gives a kibble durability and keeps fines down. Amylopectin is branched, hydrates more easily, and gives you expansion and a lighter piece with a weaker structure. The ratio between them is doing more work in your product than almost anything else in the formula.

Here is the part that catches people. That ratio is not a fixed property of the ingredient. It moves with crop variety, with growing region, and with the season the grain came out of. It moves enough that published figures for the same crop disagree with each other, which should tell you something about how much confidence to place in a single number for it. A new crop year, or a mill quietly shifting where it sources, moves the ratio in your formula without moving anything on the paperwork.

The certificate of analysis will not show it. Moisture, protein, ash and fiber can all be identical while the functional behavior of the starch has shifted underneath them.

So the trade-off is not one you chose. It arrived. And the fix depends on which direction it moved. If the starch got more amylopectin-rich you have gained expansion and lost durability, and you can trade some back with a small addition of a higher-amylose source. If it went the other way you are under-expanding, and the answer is more likely process than formula.

What makes this frustrating is that it looks like a process problem for weeks before anyone questions an ingredient nobody changed.

If you have a formula that drifted without anyone touching it, when did you last check the amylose content of the starch source rather than the proximate analysis?

Related: Starches