The count was short
Updated
On this page9
How it happens
A target fill weight is set from one batch of capsules and used for the next. The new batch is slightly heavier per capsule, the same weight goes in, and the bottle holds fewer than the label says.
Why capsule weight moves
Fill density varies with the powder blend and its moisture, shell weight varies between lots, and a hygroscopic blend gains weight in a humid room. None of that is a fault, and all of it moves the arithmetic.
The fix is a per-batch unit weight
Weigh a sample of the actual capsules, derive the target, and verify by hand counting at intervals. 21 CFR part 111 expects in-process controls over packaging and labeling operations, and this is the obvious one.
It interacts with the iron calculation
A count that drifts upward raises the elemental iron in the package. If you are near 250 mg, process drift can carry you across a threshold you calculated as clear.
Which argues for headroom on both sides
Enough above the labeled count that nobody is short, and enough below the iron threshold that drift does not cross it. Those two pull in the same direction only if the threshold is comfortably far away.
Softgels are the awkward case
They vary more than tablets and they deform, so both weighing and counting are less precise. More frequent verification is the honest answer rather than a tighter target.
Record it, per batch, with the date
Unit weight, target, sampled counts, and who checked. It is the record that answers both a short-count complaint and a question about the iron total, and it is one sheet.
It is the failure a customer can audit
Nobody at home can measure a moisture barrier, and anybody can count capsules. That asymmetry is why the count control earns more attention than its difficulty suggests.