The scoop is wrong

This is the failure customers never report, because they have no way to detect it. It is also the largest measured error anywhere in this property.

How wrong it can be

Measured scoop accuracy across several powders produced error rates from plus or minus 18 percent to nearly plus or minus 40 percent.

A serving declared at 5 g can plausibly be 3 g or 7 g. The customer believes they are taking a defined amount and they are taking a range.

Four sources

Volume against weight. The scoop measures volume, the label declares weight, and bulk density converts between them. It is not a constant.

Technique. Leveled, heaped and tamped are three different masses from the same scoop, and no instruction reliably produces one.

Moisture. A powder that has taken up water flows and packs differently, and moisture uptake degrades dosing accuracy directly.

Settling. Powder compacts in transit and further as the tub is handled, so the same scoop holds more mass later than it did at the start.

Why it gets worse down the tub

All four sources point the same way over time. The powder compacts, it takes up moisture through repeated opening, and the customer develops a habit that may not be leveling.

So the last third of a tub reliably delivers a different dose from the first third, and if the blend has also segregated, a different composition as well.

Reducing it

Declare the serving in grams alongside the scoop, so there is a true number available. Calibrate the scoop against your own finished powder rather than a nominal capacity, and re-calibrate whenever the formula or grind changes.

Specify leveled in the directions. Control moisture, which addresses this and three other failures at once. And for anything potent enough that the error matters, say plainly that weighing is more accurate.

References