Who Signs Off a Batch at a Nutraceutical Contract Manufacturer

The pallet is shrink-wrapped, labeled and invoiced. It still does not ship, because nobody can produce the record showing who approved and released the batch. Your relaunch date now belongs to someone else's filing system, and every week spent hunting an email chain is a week of warehouse cost and missed reorders.

Batch approval is a quality control act, not a line item you attach to a purchase order. Under 21 CFR Part 111 the review and release decisions sit with quality control personnel, and the question worth asking is narrower than whether your contractor has QA: can you name the person who approved and released each of your batches? What follows is a staged test for the first production run, from day one through the repeat batch that follows it.

What a missing approval signature costs you

The expensive failure is not the rejected batch. It is the batch nobody can release. Picture a finished SKU sitting in a contractor's warehouse, labeled, tested and ready, held because the release decision was never signed or cannot be located. You are paying to store goods you cannot sell, and your launch date is being set by how fast a third party can pull a file.

One document converts finished goods into shippable goods: approving and releasing, or rejecting, each finished batch for distribution, including any reprocessed finished batch, at §111.123(a)(8). Everything upstream exists to make that decision defensible, and the obligation behind it is broad — quality control personnel must ensure that manufacturing, packaging, labeling and holding operations ensure the quality of the supplement and that it is packaged and labeled as specified in the master manufacturing record.

Two failure modes burn real time. In the first, the release exists but the file behind it does not, so a retailer's question or a records request cannot be answered from what you hold. In the second, the release is signed by someone whose role in the chain nobody can describe. Both surface at the worst moment: after the goods are made and before they are sold.

Four approval acts to request on day one

Ask for four things by name on the first call. A request phrased as “send your QA documentation” returns a certificate and a quality manual. The same request phrased as four specific acts returns dated signatures tied to your product.

First, review and approval of all master manufacturing records and all modifications to them, at §111.123(a)(1). This is the recipe-level document. If it changed and the change was never approved, every batch made under the changed version inherits that gap.

Second, review and approval of all batch production-related records, plus review of all monitoring required under subpart E, at §111.123(a)(2)-(3). That covers the in-process checks logged during production, signed off against what the master record specified, and it is the layer most often summarized verbally rather than shown.

Third, the release decision itself at §111.123(a)(8), including any reprocessed finished batch. Fourth, the §111.105 obligation that quality control personnel ensure packaging and labeling match the master manufacturing record, which is the act that catches the failure everyone underestimates: a correct formula in a wrong label.

Whose QC name is on the approval

The regulation names quality control personnel as the actor. It does not name a job title at a specific company, and that silence is where brands get hurt.

Here is my standing preference: I would rather sign with a plant whose QC approver I can name, call and put in writing than with one that promises QA covers everything. A certificate proves a system exists somewhere in the building. A signature proves a person looked at this batch, on this date, against this specification.

By the end of the first week you should be able to write one sentence and have each name confirmed by the company that employs them: this person approves master record changes, this person reviews batch records and monitoring, this person releases finished batches. If the answer is a department rather than a person, you have learned something useful before the first purchase order rather than after the third.

Week three tests a master record change

Three weeks in, test whether change approvals are real. Take one product's master manufacturing record, request its modification log, and for each change ask for three data points: the approval date, the approver, and the version that actually went to the floor. Compare each entry against the requirement that all master manufacturing records and all modifications be reviewed and approved.

The test passes when the approval date precedes the first batch made under the new version, and when the version number on the batch record matches an approved version in the log. It fails when the log carries dates but no names, or names but no version reference. That pattern usually means the log was assembled for your request rather than kept as part of production.

Suppose the change was an excipient level or a new ingredient supplier. If the modification was never approved, you cannot see it in the finished goods. You can only see it in the file.

One month later, trace one batch backwards

One month after the first batch, run the trace yourself. Pick a single finished lot and walk backwards: start at the release decision, step to the batch production record and the subpart E monitoring review, then to the master record version the batch was made against.

The trace holds when the three documents point at each other by date and version. The release cites the batch, the batch record cites the master version, and that version carries an approval date earlier than production. It breaks when the release is signed but the batch record is not, or when the batch record cites a version number that never appears in the log.

Run it on a reworked lot if one exists. Reworked material is where the trail most often skips a step, because a reprocessed finished batch needs a release decision of its own rather than a footnote attached to the original run.

Repeat batches need the same approval sheet

Repeat batches are where the trail decays. The first run gets attention because everyone is watching it. The fourth run gets a signature and a shrug.

By the end of the first run, build a one-page per-batch approval sheet that carries the four acts forward: master record version and its approval date, batch record and monitoring reviewed, release decision signed, and confirmation from quality control that the packaging and label were checked against the approved master record. One page, one lot, four signatures. Design it once and the fifth run no longer requires re-learning the file.

Two items need their own slot on that sheet rather than an inherited checkmark. Any formula or process modification needs its own approval before the next batch starts, and you need to know whether rework happened at all before you can check whether it was covered.

What a contract can and cannot take over

Take each of the four acts and write three columns for your own product: who performs it, who signs it, and where the signed record lives. Do it for the SKU in front of you, not in the abstract.

The exercise usually shows that manufacturing, monitoring and the physical release happen inside the plant, while the master record and the specification your product was designed against originate on your side or with a formulator. The quality obligation sits with quality control personnel, and the citations do not say which company employs them, which is why the three-column map rather than the QA clause in the contract tells you whether anything is documented.

What the exercise does tell you is whether the handoff is documented anywhere other than a sales conversation. If the answer to where the signed release lives is that the plant keeps it, your next questions are how fast you receive a copy and whether your product name appears on it.

Who bears legal responsibility for a defective batch is a contract question for counsel. Whether you can produce a dated approval trail for a lot is an operational question you can settle this week.

What would overturn this verdict

Two conditions would change the conclusion. If you already hold dated approvals naming the person responsible for each act on your batches — master record, batch record and monitoring, release, and labeling — then the concern is answered rather than avoided, and the remaining work is keeping it current through staff turnover. Second, if your operation falls outside US dietary supplement manufacturing under 21 CFR Part 111, these citations do not govern you, and the equivalent question is whatever your own regime requires.

The limits deserve to be stated plainly. These are regulatory texts from the eCFR snapshot of October 19, 2026, not an audit. They establish nothing about any specific contractor's compliance, price, lead time, capacity or the clinical performance of any finished product. A contractor can be fully compliant and still be a poor fit for your volume, and that judgement comes from your own qualification work.

The objection readers carry out of all this is usually some version of: our contract says the manufacturer handles all QA, so we are covered. A contract can allocate who does the work. It cannot make a signature appear on a document you do not hold.

So request the four approval acts on your first batch, in writing, with names and dates, and put your own named reviewer in the file. If the signature is missing when the pallet is ready, treat that as a procurement answer — a fact about how this supplier operates — rather than a paperwork delay you can wait out. My preference does not shift with the setting: a name you can call beats a department you cannot.