Accepted answer
Twelve months is 52 weeks, so the budget is set by lot turnover, not by the price of a sterility test. Take one lot a quarter as the low case: 4 lots a year, so a test-every-lot policy is 4 assays and a test-every-third-lot policy is 2 once you round up. Take one lot a month as the high case: 12 lots, and the same two policies are 12 assays and 4. The spread between the cheapest and the dearest defensible policy is therefore about a factor of six across the same 52 weeks. Choose the policy before the first result. One chosen after a disappointing figure is a reaction to that figure, and it will not survive the second one. Then spend it where it changes a decision: over a year, one content assay on each new lot tells you more than four purity figures on the same lot, because purity and content are independent and only one of them changes your arithmetic.
The practical consequence is that spot-testing one vial from a new supplier is better than assuming they are all the same.
Stratified sampling — testing one vial from the top, one from the middle, and one from the bottom of a shipment — is cheap insurance against segregation.
On the detail: the sample size determination requires choosing a confidence level and an acceptable error rate, and the smaller the error rate you want, the larger your sample must be.
Sampling plans for pharmaceutical manufacturing are defined in ISO 2859 and ANSI Z1.4, and they are based on statistical sampling theory.
I would treat a "complies with" statement without sampling details as a claim rather than as evidence.
The practical summary: a lot number without a sampling statement is a lot number without meaning.