Do I have to fully validate a compendial method, or is verification enough?

This is United States federal law, and the first thing to settle is whose. The sentence this whole argument turns on sits in the drug good manufacturing practice regulations, and that part carries the minimum current good manufacturing practice for the preparation of drug products, excluding positron emission tomography drugs and medical gases, for administration to humans or animals (21 CFR 211.1(a)). If you make a dietary supplement, a conventional food or a cosmetic, you are not preparing a drug product, and the answer people quote at you out of that part is not your answer. Yours is further down this page, and it is not the same answer.

What this page does not reach. It settles which route the rule asks for — a full validation, or a verification — and what each one has to show. It does not tell you how to run either study. It does not cover moving a method from one laboratory to another as a subject of its own. And it does not cover process validation, which proves the process makes good product rather than that the test measures it. Those are separate questions and this page answers none of them.

Three words carry the weight below, so they are settled first.

• Compendial, for this purpose, is the rule’s own list: a method in the current revision of the United States Pharmacopeia, the National Formulary, the AOAC INTERNATIONAL Book of Methods, or in other recognized standard references (21 CFR 211.194(a)(2)). AOAC sits in that list beside USP. A method is not compendial because it looks official, because it has a number, or because a supplier called it a standard method.
• Validation, in the regulation’s own words, is the accuracy, sensitivity, specificity and reproducibility of the test methods a firm uses being established and documented (21 CFR 211.165(e)). Four characteristics, named. That is narrower than the list the industry usually recites.
• Verification is not defined in the regulation at all. What the regulation asks for is written as an act rather than as a term: the suitability of all testing methods used shall be verified under actual conditions of use (21 CFR 211.194(a)(2)).

The pharmacopeia publishes its own general chapters on both, numbered 1225 for validation of compendial procedures and 1226 for verification of them, and FDA names both in its guidance on analytical procedures (FDA, Analytical Procedures and Methods Validation for Drugs and Biologics, July 2015). Those chapters are the pharmacopeia’s own text. They are not reproduced here and this page makes no claim about what they require; read them at the source.

What the statute says about a compendium is narrower than it is usually reported to be, and it is about test methods rather than about general chapters. Where a drug purports to be or is represented as one whose name is recognized in an official compendium, whether its strength, quality or purity meets that compendium’s standard shall be determined in accordance with the tests or methods of assay set forth in that compendium — except where the compendium prescribes none, or where the Secretary judges those prescribed insufficient, in which case the Secretary promulgates regulations prescribing the tests by which the determination is made (21 U.S.C. 351(b)). For a supplement the hook is a different one and it turns on what you said about your own product, which is dealt with below.

If you make a drug: the rule, read to the end

One paragraph of the regulation does three separate jobs, and the argument survives because people stop after the second.

First, laboratory records must include a statement of each method used in testing the sample, and that statement shall indicate the location of data establishing that the methods used meet proper standards of accuracy and reliability as applied to the product tested (21 CFR 211.194(a)(2)). That is the data package, and note the last five words.

Second, in parentheses inside the rule itself: if the method employed is in the current revision of the United States Pharmacopeia, National Formulary, AOAC INTERNATIONAL Book of Methods, or in other recognized standard references, or is detailed in an approved new drug application and the referenced method is not modified, a statement indicating the method and reference will suffice (21 CFR 211.194(a)(2)). That is the carve-out, and it is real. It relieves you of generating and locating the accuracy-and-reliability data. It is what people mean when they say a compendial method does not need validating.

Third, and this sentence carries no carve-out of its own: the suitability of all testing methods used shall be verified under actual conditions of use (21 CFR 211.194(a)(2)). All of them. Read that word against the sentence before it. The compendial route removes the data package. It does not remove the verification, and the verification is the part that costs you laboratory time.

The regulation then settles the terminology argument itself. Accuracy, sensitivity, specificity and reproducibility shall be established and documented, and such validation and documentation may be accomplished in accordance with the laboratory-records provision above (21 CFR 211.165(e)). Read the verb. The rule calls that route validation and offers it as a permitted way of satisfying the requirement, not as the only one. So the honest answer to the question as buyers ask it is not that verification is a cheaper substitute for validation. It is that the rule lets the compendial route stand as the validation for an unmodified compendial method, and the verification under actual conditions of use is the part of that route you still have to perform.

FDA says something close to this in plainer words in its own guidance, and it matters who that guidance is addressed to: its recommendations apply to drug substances and drug products covered in new drug applications, abbreviated new drug applications, biologics license applications, and supplements to those applications. Inside that frame it says compendial methods are verified rather than validated; that the suitability of a procedure from USP/NF, the AOAC Official Methods of Analysis or other recognized standard references should be verified under actual conditions of use; and that information demonstrating USP/NF procedures are suitable for the drug product or drug substance should be included in the submission and generated under a verification protocol carrying the compendial methodology to be verified with predetermined acceptance criteria and the details of the methodology (FDA, Analytical Procedures and Methods Validation for Drugs and Biologics, July 2015, section VI). That guidance states on its face that it is nonbinding and creates no rights. The regulation is the binding text and it reaches every finished pharmaceutical. The guidance tells you how the agency reads the subject in an application it is reviewing, and if you hold no such application it is a useful account rather than an instruction to you.

Which route your method is on

What you are running	What the rule asks of you	Where it says so
A compendial method you have not modified	Verify its suitability under your actual conditions of use, and record it. A statement of the method and its reference stands in for the accuracy-and-reliability data.	21 CFR 211.194(a)(2)
A compendial method you changed	Complete records of the modification, the reason for it, and data verifying that the modification produced results at least as accurate and reliable for the material being tested as the established method.	21 CFR 211.194(b)
A method that is not compendial	Accuracy, sensitivity, specificity and reproducibility established and documented.	21 CFR 211.165(e)
Any testing method at all, whatever its source	Suitability verified under actual conditions of use.	21 CFR 211.194(a)(2)

Behind all four sits the general requirement the four are instances of: laboratory controls shall include the establishment of scientifically sound and appropriate specifications, standards, sampling plans and test procedures (21 CFR 211.160(b)). A method can satisfy the paperwork of the table above and still fail that sentence.

Two things the carve-out does not do

It does not travel to your product

The first sentence of the provision ties the data to the thing tested: accuracy and reliability as applied to the product tested (21 CFR 211.194(a)(2)). FDA makes the same point about carrying a method across products, in that same application context: to apply an analytical method to a different drug product, appropriate validation or verification studies for compendial procedures with the matrix of the new product should be considered (FDA, Analytical Procedures and Methods Validation for Drugs and Biologics, July 2015, section II).

This is the failure underneath the question, and it is not a paperwork failure. A method that carries a recognized name gets trusted to detect something, and the trust is placed in where the method came from rather than in evidence from the matrix it is being run on. Excipients that were not in the monograph product co-elute. A botanical extract quenches the signal. A softgel shell interferes with the recovery step. The result comes back clean and sits quietly alongside the very thing it was trusted to rule out, and nothing about the number tells you that.

It does not survive a change you made

The carve-out is conditional on the referenced method not being modified. Change it and you are outside it, and a different requirement attaches: complete records shall be maintained of any modification of an established method, including the reason for the modification and data to verify that the modification produced results at least as accurate and reliable for the material being tested as the established method (21 CFR 211.194(b)).

The requirement is not the trap. The trap is that the change is rarely recognized as one. A column swapped for the one on the shelf. A flow rate nudged to bring the run time down. An extraction time extended because recovery was low. Each is an alteration to a procedure, each is easier to make than to write up, and none of them feels like a change control on the afternoon it happens. Nobody sets out to leave a method unqualified. They set out to get the run finished, and the method quietly stops being the method the reference describes.

What the inspection record shows

FDA publishes the observations its investigators write up at inspections, each tied to the provision it was written under. Read on 18 August 2026, counting only observations whose cited provision is exactly the laboratory-records paragraph or the supplement laboratory-methods paragraph:

Provision cited	Observations	Establishments	Years covered
21 CFR 211.194(a)(2) — drug program	142	123	FY2009 through FY2026, present in every one of the eighteen years
21 CFR 111.320(a) — food program	61	53	FY2009 through FY2026, present in every year but one

Of the 142 drug-side observations, 120 carry text about verification of the suitability of the testing methods, across 106 establishments; 110 of the 142 say in terms that the suitability was not verified, or not performed, under actual conditions of use. The rest of the 142 are about the laboratory record itself rather than the verification: the method not stated, the reference not given, the location of the accuracy-and-reliability data missing. All 61 on the supplement side carry a single sentence: you did not verify that the laboratory examination and testing methodologies are appropriate for their intended use.

What that counts, and what it does not. These are observations written at the time of an inspection, not final agency determinations, and each one records that a firm did not do something. Set against roughly 41,600 drug-program and 154,800 food-program observations in the same data, this is a thin, steady line rather than a common finding. It is evidence that the failure is real and that it keeps happening year after year. It is not evidence of how many firms were inspected, of what share of them verify properly, or of whether the verifications that do exist are any good. A count of a thing being absent says nothing about the quality of the ones that are present.

If you make a dietary supplement

The drug part is not addressed to you. The part that is applies to you, except as its own next paragraph provides, if you manufacture, package, label or hold a dietary supplement (21 CFR 111.1(a)); that exception lifts the holding requirements from a retail establishment holding supplements for the sole purpose of direct retail sale to individual consumers (21 CFR 111.1(b)).

The laboratory-method duty sits in two places in that part, and you need both.

The laboratory-methods section says you must verify that the laboratory examination and testing methodologies are appropriate for their intended use (21 CFR 111.320(a)), and that you must identify and use an appropriate scientifically valid method for each established specification for which testing or examination is required to determine whether the specification is met (21 CFR 111.320(b)).

The specification-testing section says it again from the other end, and adds something the first one does not. You must ensure that the tests and examinations you use to determine whether the specifications are met are appropriate, scientifically valid methods (21 CFR 111.75(h)(1)), and those tests and examinations must include at least one of gross organoleptic analysis, macroscopic analysis, microscopic analysis, chemical analysis, or another scientifically valid method (21 CFR 111.75(h)(2)). That last list is the floor. A supplement maker who reads only the laboratory-methods section will miss it.

Now read what is not there. There is no compendial carve-out, so a USP method buys you no shortcut on the face of the rule. There is no list of validation characteristics either, so nothing in the part demands a full validation package of the kind a drug submission carries. The two tests are appropriateness for the intended use, and scientific validity against the specification you set. If you have not settled what the specification is, the method question cannot be answered yet, and that is a question of its own.

Where a compendium becomes binding on you is in the statute, and it turns on what you claimed. A supplement is misbranded if it is covered by the specifications of an official compendium, is represented as conforming to those specifications, and fails to so conform (21 U.S.C. 343(s)(2)(D)(i), and its clauses (ii) and (iii)). Print USP on the label and you have taken on the compendium’s specifications by your own hand.

And where no compendium covers the product, the statute reaches the method directly. A supplement is misbranded if it fails to meet the quality, including tablet or capsule disintegration, purity, or compositional specifications, based on validated assay or other appropriate methods, that the supplement is represented to meet (21 U.S.C. 343(s)(2)(E)(ii)(II)). That is where the word validated actually lands on a supplement maker. Not on the pedigree of the method, but on whether the assay standing behind the claim you printed will carry it.

If you hold a supplement good manufacturing practice certification, the scheme is more explicit than the regulation, and its own verbs are worth separating. It requires that test methods and examinations be identified, scientifically valid, and verified as appropriate for their intended use. It then, in the weaker verb it reserves for a recommendation, says compendial methods should be verified with predetermined acceptance criteria to show they are competently executed and suitable for use, and that documentation of that verification should be available and suitable (NSF/ANSI 455-2, clause 4.6.13 and its subclauses). The regulation states the verification duty in general terms. The scheme states it against compendial methods by name, and asks for the paperwork.

If you make a cosmetic

No federal rule tells you which analytical method to use, or how to qualify it. That is not the same as saying nothing federal reaches your testing, and the difference is worth holding on to before you spend money answering a question the law has not asked you.

On manufacturing practice, the statute makes a cosmetic adulterated if it has been manufactured or processed under conditions that do not meet the good manufacturing practice requirements of the section that governs them (21 U.S.C. 361(f)), and that section directs the Secretary to establish those practices by regulation (21 U.S.C. 364b(a)). The duty runs through regulations, not through the statute on its own. The same section set a schedule for them: a proposed rule within two years of 29 December 2022, and a final rule within three (21 U.S.C. 364b(c)).

Read on 18 August 2026, the cosmetics subchapter of the Code of Federal Regulations carries a general part, a labeling part, two voluntary filing programs and a warning-statements part, and the parts after those are reserved. There is no good manufacturing practice part in it. So this is a pending position rather than a settled one, and it is exactly the kind of statement that stops being true without anyone telling you. Check it at the source before you rely on it, including when you read it here.

One federal obligation does reach your testing, and it arrives from a different direction. A cosmetic product is adulterated if the product, including each ingredient in it, does not have adequate substantiation for safety (21 U.S.C. 361(g)); a responsible person must ensure, and keep records supporting, that there is adequate substantiation of safety (21 U.S.C. 364d(a)); and adequate substantiation of safety means tests or studies, research, analyses, or other evidence or information considered, among experts qualified by scientific training and experience, sufficient to support a reasonable certainty that the product is safe (21 U.S.C. 364d(c)(1)). That can rest on testing, and it names no method. It tells you the evidence has to satisfy qualified experts. It does not tell you to validate or verify anything.

So on the method question specifically, what binds you is your customer’s specification, and where you are certified, the scheme’s. A cosmetics good manufacturing practice scheme asks that scientifically valid test methods be used for components, packaging materials, in-process materials and final products; that compendial methods may be used; that where methods are validated, the USP or ICH criteria for method validation should be followed; and that method transfer protocols with predetermined acceptance criteria be defined and followed whenever a method developed and validated at a different laboratory is taken up (NSF/ANSI 455-3, clause 4.6.3 and its subclauses). Read the verbs there. Compendial methods may be used, and validation criteria should be followed. That is a different instrument from a regulation, and the thing that makes it bite is that you agreed to it.

If you make an over-the-counter drug

An over-the-counter drug product is a drug product, so everything above about the drug regulation reaches you on the same terms as any other finished pharmaceutical. The FDA guidance quoted above does not, unless you hold one of the applications it addresses; if you market under a monograph, that guidance is an account of the agency’s reading rather than a set of recommendations aimed at you.

One boundary is written into the scope of the part and is worth reading rather than assuming: pending consideration of a proposed exemption published in the Federal Register on 29 September 1978, the requirements of the part shall not be enforced for over-the-counter drug products if the products and all their ingredients are ordinarily marketed and consumed as human foods, and the products may also fall within the legal definition of drugs by virtue of their intended use. Until further notice, the regulations in parts 110 and 117, and where applicable parts 113 through 129, are applied instead in determining whether such products are manufactured, processed, packed or held under current good manufacturing practice (21 CFR 211.1(c)). That is narrow, and note what it decides: which practice regulations you are measured under, not whether the product is adulterated on some other ground. It turns on the product and every one of its ingredients being ordinary human foods, not on where the product is sold.

If you carry an over-the-counter drug good manufacturing practice certification, the scheme is firmer than the regulation on this exact point: the USP or the ICH criteria for method validation shall be followed — the scheme offers the choice — test procedures are validated confirming accuracy, sensitivity, specificity and reproducibility, and compendial methods are verified to show that they are suitable for use with the verification documentation available (NSF/ANSI 455-4, clause 4.6.6 and its subclauses).

Where the guideline everybody cites actually applies

When somebody says full validation, they usually mean the international guideline on validating analytical procedures. Read its own scope before you decide it governs you: it applies to analytical procedures used for release and stability testing of commercial drug substances and products, and can also be applied to other analytical procedures used as part of the control strategy following a risk-based approach (ICH Q2, section 1.2). It is written to say what data should be presented in a regulatory submission.

So on a supplement, a food or a cosmetic, that guideline is a reference you may choose to work to, or that a customer may impose on you in a specification. It is not a rule reaching you on its own, and citing it in your protocol does not make it one. It is also one of the two criteria sets your certification scheme points at when it says validation criteria; the schemes name the USP or the ICH, and leave you the choice. That is how a guideline written for drug submissions ends up governing a vitamin gummy: not by law, but by contract.

What clean looks like

Four things, and two of them you can settle from this page.

1. The method’s source is documented and the method is unmodified. The reference is named, the current revision is the one you hold, and the procedure your analyst runs matches it line for line. You can check this yourself against the reference in front of you.
2. Every modification is written up. Each change carries its reason and the data showing results at least as accurate and reliable as the established method. You can check this yourself by reading your method history against your current procedure.
3. Suitability is verified under your actual conditions of use, on your matrix, and documented. A protocol with acceptance criteria set before the work, and a report against them. You cannot settle this from a page. Whether the verification you ran was deep enough for your matrix is a judgment about your data, not a reading of a rule.
4. The method can detect what the specification exists to catch. A method that passes at release and cannot see the degradation product is compliant on paper and blind in practice. You cannot settle this from a page either. It is a question about your product’s failure mode, answered from your own stability and impurity data.

If all four come back clean, you do not need anything else on this page and you do not need to buy anything.

Three things to check before you spend

1. Ask what the method was verified on, not where it came from. The useful question to your laboratory is not whether the method is a USP method. It is which product and which matrix the suitability was verified on, and whether that report exists. If the answer names a different product, you have your gap.
2. Read your own method against the reference, side by side. Not the summary, the procedure. Every difference is either a documented modification carrying its own data or an undocumented one, and there is no third option in the rule (21 CFR 211.194(b)).
3. Find out what you promised. For a dietary supplement, a compendium becomes binding through your own representation, and the statute wants all three conditions together: the supplement is covered by an official compendium’s specifications, is represented as conforming to them, and fails to so conform (21 U.S.C. 343(s)(2)(D)(i), with clauses (ii) and (iii)). That provision reaches dietary supplements and nothing else. For a cosmetic, the equivalent question is what your customer specification and your certification scheme say, because that is where a method obligation comes from at all.