Key takeaways
Match the CoA to the exact product, batch, sample stage and approved specification revision.
Check results against the limits, units and test conditions needed to interpret them.
Laboratory conformity is part of release, alongside production, packaging, deviations and market-readiness checks.
Name the release authority and escalation process in the quality agreement; a CoA is not regulatory certification.
In this article
A cosmetic certificate of analysis (CoA) reports results or conformity statements for an identified batch against specified criteria. Read it alongside the approved specification and batch record. Passing test results do not establish that the correct formula went into the correct packaging or that an authorized person released the batch for sale.
When choosing your cosmetic manufacturer, agree who can hold stock and who can authorize release. Give the quality reviewer access to the supporting records as well as the certificate.
What the CoA covers
The specification states what is acceptable. The CoA reports what the issuer found or declared for the identified batch. A certificate of conformity may simply declare compliance without listing results. Ask what the document covers rather than relying on its title.
For EU cosmetics, manufacturing must comply with good manufacturing practice under Regulation (EC) No 1223/2009, Article 8. Article 10 separately requires the product safety assessment and report; Article 11 requires the product information file. A batch CoA does not replace those obligations.
ISO 22716:2007 covers production, control, storage and shipment of cosmetic products. Its public abstract describes that scope, not a universal CoA template. Use the checklist below as a quality-review aid. It is not a mandatory CoA format or testing panel for all cosmetics.
For US cosmetics, FDA’s MoCRA overview explains the responsible person’s duty to maintain adequate safety-substantiation records. Passing a batch’s laboratory tests does not by itself satisfy that duty. This article addresses cosmetics, not drug-release requirements for products such as US OTC sunscreens.
Match the certificate to the stock
Before reading the results, check that the certificate identifies the stock under review. Clarify any missing link:
- Product name, product code and applicable formula or specification revision.
- Manufacturer and test laboratory, distinguishing the document issuer from the site that performed testing.
- Manufacturing batch and finished-goods lot, with a documented link if filling created separate lots.
- Sample identity and stage: raw material, bulk, filled product or retained sample.
- Relevant manufacturing, sampling, testing and certificate issue dates.
- Authorized issue or approval, document version and any replacement-certificate reference.
A bulk test result cannot, by itself, address contamination introduced during filling. Ask how the sampling plan addresses the manufacturing and filling stages, including split fills or multiple packaging formats.
FDA’s 2013 draft cosmetic GMP guidance, in its records and laboratory-controls sections, recommends representative sampling and records linking ingredients, production steps, finished-product lots and acceptance status. These are examples from historical, nonbinding draft guidance, not enforceable US rules.
Read the result with its method
Compare every required attribute with the specification approved for that batch. Do not accept a supplier’s wider default limit in place of the agreed product limit without an authorized change review.
| CoA field | Review question | Reason to pause |
|---|---|---|
| Acceptance criterion | Is this the approved range or requirement? | Missing limits or an unexplained revision |
| Result and unit | Is it a measured number, qualitative result or declaration? | Blank entry, incompatible units or ambiguous “pass” |
| Method and conditions | Does it reference the agreed method and relevant conditions? | Viscosity without the conditions needed to interpret it |
| Sample and date | Was the intended batch stage sampled at an appropriate time? | An unrelated bulk lot or implausible chronology |
| Reporting scope | Was the attribute tested on this batch or supported another way? | Periodic testing presented as a fresh batch result |
For viscosity, ask for the relevant temperature, instrument setup, spindle and speed or a method reference containing them. For pH, check whether the method measures the product directly or a specified dilution. The details needed will depend on the formulation and agreed method.
For microbiology, read both the count and any specified-organism results. “Not detected” means not detected under that method’s conditions and sample quantity; it does not mean absolute absence. A reported result below a quantification or detection threshold is not automatically zero.
The laboratory’s microbial limits test and a preservation challenge test answer different questions. Annex I of the EU Cosmetics Regulation addresses microbiological specifications and challenge-test results within the safety information; the Commission’s section 3.3 guidance discusses scientifically justified testing approaches for different product risks. See the distinction in our microbial limits and challenge-testing guide.
Hypothetical example: passing numbers, incomplete release
Imagine a moisturizer batch identified as EX-026. The following entries are invented for this example. They are neither recommended limits nor evidence from an actual manufacturer.
| Attribute | Hypothetical approved criterion | Hypothetical reported result | Review |
|---|---|---|---|
| Appearance | Matches approved reference | Conforms | Confirm the reference and inspection method |
| pH | 5.0 to 5.6 by method PH-01 | 5.3 | Within the illustrative range |
| Viscosity | 8,000 to 12,000 mPa·s by method V-02 | 10,200 mPa·s; method blank | Obtain the method reference and supporting conditions |
| Microbiological panel | Product-specific approved requirements | Pending | The required results are incomplete |
The pH result meets the illustrative criterion. The viscosity number is within range, but the missing method reference needs clarification. Release must also wait for the required microbiology results.
Suppose the filling record then identifies bulk EX-025 rather than EX-026. Even a complete, passing CoA for EX-026 would not resolve that mismatch. Keep the affected stock on hold and investigate which bulk entered the filled units. Require traceable evidence and an approved conclusion rather than an informal assurance that the operator made a typo.
Review the rest of the release package
Use a checklist that records the evidence reference, reviewer, date and disposition. Before authorizing sale, confirm that:
- The product, formula, specification and batch identities agree across the CoA, production and filling records.
- Required tests are complete, meet the applicable criteria and use the agreed methods; any testing exemption has an approved basis.
- Production records account for approved ingredient lots, processing, adjustments, relevant hold times and yields.
- Packaging and label versions match the approved product, with recorded line checks and correct lot coding.
- Deviations, unusual results, rework and discrepancies have documented assessments and authorized dispositions.
- Relevant stability, packaging and preservation evidence still applies to the formula and process used.
- Market-readiness checks are complete, including applicable safety documentation, labelling and notifications.
- The authorized person records the release decision and the warehouse receives an unambiguous stock-status instruction.
The CoA can reference supporting reports rather than reproducing them all. Your team still needs access to the evidence required by its review procedure. After a manufacturing change, check whether existing stability and packaging compatibility testing remains relevant.
Decide who may release, hold or reject
Name the authorized functions in the quality agreement and procedures: who approves results, investigates discrepancies, accepts deviations, authorizes shipment and authorizes sale. Separate those decisions if the brand requires review after the manufacturer’s own release. A purchase-order deadline is not authorization to release stock.
These operational assignments do not transfer statutory obligations. In the EU, Article 5 makes the responsible person accountable for ensuring compliance with the listed requirements, including GMP and the safety assessment. A contract can allocate tasks and access to records, but a CoA is not approval from a regulatory authority.
If a result fails or a required record is missing, hold the affected stock and follow the investigation procedure. Preserve the original result and document any retesting rationale. Repeated testing until a passing number appears does not explain the original failure. Agree how the manufacturer alerts the brand and who assesses potential effects on other batches.

Sources and references
Frequently asked questions
Does every cosmetic batch need a challenge test?
Does every cosmetic batch need a challenge test?
Preservation challenge testing is not automatically required on every manufactured batch. It evaluates preservation performance under defined conditions; routine microbial checks assess the sample’s microbiological quality. Agree the testing programme and any justified exceptions with the laboratory and assessor for your product.
Can a CoA contain only “pass” results?
Can a CoA contain only “pass” results?
Some qualitative attributes can reasonably report conformity to a defined criterion. For quantitative attributes, request the numerical results and units your approved review requires. Establish whether a statement reflects a test on this batch, periodic testing or another basis.
Does an ISO certificate release the batch?
Does an ISO certificate release the batch?
No. A manufacturer’s quality-system certificate does not provide this batch’s results, resolve its deviations or record its release decision. Review the batch evidence and follow the authority assigned in the quality agreement.

