In fiscal year 2023, federal investigators carried out 501 inspections of dietary supplement facilities. The single most common thing they wrote down, in 78 of those inspections, was not a dirty tank or a contaminated batch. It was that the company had never established what its own product was supposed to be.
That is the citation at the top of the list: failure to establish product specifications for the finished batch, which the trade analysis of the fiscal 2023 data identifies as 21 CFR 111.70(e). It was the most common observation in fiscal 2017 and fiscal 2018 too. Fiscal 2021 is reported a little differently: three of the top five citations that year concerned specifications, and an FDA spokeswoman said finished product specifications and written quality control procedures remained the two most-cited violations. The most frequent violation of American supplement manufacturing law is not writing down what you are making.
A note about all of those counts before going further. None of them is an FDA publication. Each was compiled by a trade journal that requested the underlying inspection data and did the arithmetic itself, because the agency releases nothing supplement-specific. A later section explains why, and it is not a small point.
Now set that beside the second number. In October 2022, Cara Welch, who directs the Food and Drug Administration’s Office of Dietary Supplement Programs, said FDA runs about 500 to 600 supplement inspections in a typical year, roughly 5 percent of as many as 10,000 registered facilities. That 5 percent is a 2022 figure and it has moved since: the 800 inspections recorded for fiscal 2025, in the table below, work out to about 8 percent of the same 10,000. Either way, a plant can go years without an inspector walking through the door.
This article is about the factory and the inspection record. Companion articles cover the certifiers, the laboratory surveys of what ends up in bottles, and the wider question of what the regulatory system does and does not verify.
What the rule is
The federal supplement manufacturing rule was published as a final rule on 25 June 2007 and lives at 21 CFR Part 111. It applies to anyone who manufactures, packages, labels or holds a dietary supplement.
Part 111 runs to sixteen subparts, lettered A through P. They cover personnel, physical plant and grounds, equipment and utensils, the requirement to establish a production and process control system, quality control, components and packaging and labels, the master manufacturing record, the batch production record, laboratory operations, manufacturing operations, packaging and labelling, holding and distributing, returned products, product complaints, and records.
Read that list again and notice what dominates it. Part 111 is overwhelmingly a documentation rule. One section, 111.75, requires anybody to test anything, and its obligations are narrower than most people assume. The companion article on purity and potency reads that section line by line.
The section that generates the citations is 111.70. It requires a manufacturer to establish product specifications for the identity, purity, strength and composition of each finished batch, and to set limits on contamination that could adulterate it. It also requires an identity specification and component specifications for the ingredients going in.
Read the verb. Establish. The rule requires a firm to decide and write down what its product is supposed to be. It does not tell the firm what the answer should be. It sets no values, no minimum potency for any ingredient, no ceiling on any contaminant. The company picks its own targets and is then measured against them.
That is why the top citation is what it is. A firm cited under 111.70(e) has not failed a test. It has not got to the point of having a test to fail.
It helps to picture what a specification is. For a 500 mg turmeric capsule it would state what the botanical is and how its identity is confirmed, how much curcuminoid content a finished capsule must contain, what the capsule may not contain, and the maximum permissible lead, cadmium, microbial count and solvent residue. Without that document, a laboratory result is a number with nothing to compare it to, and an investigator has nothing to check the batch record against.
The gap this leaves is visible from outside the agency too. The research group that has published most of the recent laboratory surveys of supplement content noted in JAMA Network Open in 2023 that FDA inspections have repeatedly found manufacturers failing to comply with basic manufacturing standards, naming the establishment of identity, purity and composition of the final product as the example. The citation data and the independent literature describe the same hole.
The same finding, four years running
The pattern is remarkably stable across the years for which data has been released.
- Fiscal 2017: failure to establish finished product specifications appeared in 89 inspections, about 24 percent of firms that received a Form 483.
- Fiscal 2018: the same citation, 75 inspections, about 24 percent.
- Fiscal 2021: the same citation, 59 inspections, about 40 percent.
- Fiscal 2023: the same citation, 78 inspections, close to 33 percent.
The rest of the fiscal 2023 top five tells the same story in different words. Second was failure to establish component specifications, 47 inspections. Third was failure to establish an identity specification, 43 inspections. Fourth was failure to establish and follow written quality control procedures, about 17.5 percent of cited firms. Fifth was failure to establish and follow written procedures for holding and distributing, 41 inspections, 17.1 percent.
Four of the five are about writing something down.
What that means for a reader is less damning than it sounds and more useful. These are not findings that a product was contaminated. They are findings that the system meant to catch contamination was never built. A plant with no finished-product specification has no way of knowing whether a batch is good, because it has not defined good.
The numbers, and why they are hard to come by
A figure that needs re-dating
A number travels through consumer writing about this subject that needs retiring, or at least re-dating: the claim that 70 percent of supplement facilities have manufacturing violations.
Here is where it comes from. In fiscal 2012, FDA conducted 361 supplement inspections and issued 253 Form 483s, an average of 7.2 observations each. A trade journal obtained that data under the Freedom of Information Act and divided one number by the other. Seventy percent was a publication’s arithmetic on a fourteen-year-old dataset, not a statistic FDA published.
These extracts are also less stable than a round number suggests. The same journal first reported that 72 percent of firms were cited in fiscal 2013. FDA’s FOIA office then revised the underlying data, and the corrected figure was 65 percent of 416 inspections.
The comparable rates since have roughly halved:
| Fiscal year | Inspections | Share receiving a Form 483 |
|---|---|---|
| 2012 | 361 | 70% |
| 2013 | 416 | 65% |
| 2018 | 591 | about 54% (46% received none) |
| 2021 | 278 | 53% |
| 2023 | 501 | 48% |
| 2025 | 800 | 49.1% |
Fiscal 2025 also splits by location: 720 US inspections with 47.6 percent cited, and 80 non-US inspections with 72.5 percent cited, on a much smaller base. Those figures came from a FOIA request filed by ConsumerLab, a company that sells a supplement certification seal, which is worth knowing when reading a statistic about the inadequacy of the alternative.
Two cautions belong with the whole table. A Form 483 is a list of an investigator’s observations at the end of an inspection. It is not a finding that the law was broken and it is not a penalty. And every rate here is a rate among facilities that were inspected. FDA targets its inspections rather than drawing a random sample, so none of these numbers estimates the industry.
Why all of this arrives by public records request
FDA publishes inspection observation data. It has done so since fiscal 2006, in downloadable spreadsheets broken out by program area, with tabs for biologics, drugs, devices, human tissue, radiological health, veterinary medicine, bioresearch monitoring and more.
There is no dietary supplement tab. The relevant one is labelled Foods (includes Dietary Supplements). A supplement number cannot be pulled out of it.
FDA also states two limits on its own data. The spreadsheets are not a comprehensive listing of all inspectional observations. And the counts come from FDA’s electronic inspection tools, so, in the agency’s words, “Not all Form FDA 483s are generated by these tools as some 483s are manually prepared.”
Put those together and you get the reason every supplement-specific figure in this article arrived by way of a public records request. There is no published federal series on supplement manufacturing inspections. That absence is a fact about the subject, and it is why the numbers that circulate are old, inconsistently sourced and rarely dated.
What a warning letter does
Two real letters show what an inspection produces when it goes badly.
In July 2024, FDA wrote to a California contract manufacturer. The firm had not established component specifications; the specification sheets for two botanical ingredients carried no heavy metal limits, and the letter noted those ingredients are at risk for heavy metal contamination. The firm had also not established product specifications for the strength of its finished batches. Its method for determining strength, as recorded on its own specification sheet, was “input.”
Input means how much the company put in. FDA’s response is the most illustrative line in this whole subject: measuring input may be a useful process control, but it is not an appropriate test method for determining the strength specification of those ingredients in the finished product. The company’s way of verifying how much active ingredient came out was to look at how much it had put in.
In July 2017, FDA wrote to a Utah supplement maker about inadequate finished-batch specifications, mishandled returns, and an absence of written procedures for holding and distributing. On that last point the letter adds a sentence that does more work than the citation itself: the firm had been made aware of the deficiency during an inspection in February 2013. FDA came back in October 2016, three years and eight months on, and the firm said it still had no such written procedure.
Then there is the question of what happens after the letter arrives. Researchers writing in JAMA in 2022 went back to 31 products that FDA had named in warning letters over prohibited stimulants, bought the ones still on sale and analysed them. The companion article on contamination sets out what they found.
The letter is correspondence, not an order. It carries no recall, no seizure and no withdrawal, and across that set of 31 it took exactly one product off the shelf.
The name on the bottle is usually not the factory
Most consumers assume the brand on the label made the product. Often it did not.
A large part of the industry runs on own-label distribution: a brand owner designs and sells a product, and a contract manufacturer with no name on the package makes it.
FDA’s position on who is responsible is stated plainly in its own correspondence. In a 2011 warning letter to a brand owner, the agency wrote that an own-label dietary supplement distributor contracting with other manufacturers is itself considered a manufacturer of those supplements, and as such has ultimate responsibility for the products it introduces into interstate commerce. Contracting out the work does not contract out the obligation.
How much of the market works this way is a question nobody can answer with a citable number. The two trade accounts of the subject both decline to estimate it, and the only percentages on offer come from paywalled market research that cannot be checked. The structure is clear even without a figure: brand owners who make nothing, contract manufacturers who appear on no label, and a legal duty that sits with the name you can see.
Two consequences follow for a reader. A brand’s manufacturing claim may describe a building the brand does not own. And when one contract manufacturer gets something wrong, the error propagates across every brand it fills, which is why recalls in this industry so often involve a list of unrelated names at once.
What compliance tells you, and what it cannot
FDA’s own description of the manufacturing rule is precise about its purpose. It requires activities “to ensure that a dietary supplement contains what it is labeled to contain,” and that the product is not contaminated with harmful or undesirable substances such as pesticides, heavy metals or other impurities. The rule is framed entirely around identity, purity, quality, strength, composition and contamination.
The words efficacy and effectiveness do not appear on that page. They do appear on FDA’s consumer page about supplements, in a different context: under the law that governs supplements, FDA is not authorized to approve them for safety and effectiveness before they are marketed.
So a facility can be fully compliant with 21 CFR Part 111, making a product that contains exactly what its label declares, free of contaminants, documented end to end, and none of that bears on whether the product does anything for anyone. Good manufacturing practice is a statement about consistency, not about benefit. Another article in this library handles the evidence question directly.
It is also worth knowing where the supplement rule sits relative to the drug rule, because the two are often spoken of as equivalents. A summary of the differences prepared by a regulatory consultancy and hosted by the Consumer Healthcare Products Association describes Part 111 as generally less stringent and more flexible than the drug manufacturing rules at 21 CFR Part 211, with a few areas where it is more explicit. Two differences stand out. Drug manufacturers are required to test each active pharmaceutical ingredient in each finished batch, while supplement manufacturers have the option of testing a subset of finished batches under a statistical sampling plan. And there is no requirement in Part 111 for a supplement manufacturer to formally validate its manufacturing process, a step that is mandatory for drugs. Validation is the exercise of proving that a process reliably produces the same result; without it, consistency is an expectation rather than a demonstrated property.
That comparison is a trade-association-hosted document by a consultant who sells compliance services, which is worth saying out loud. The empirical check on its last point is the galantamine study described in the companion article on purity and potency, where the same molecule measured 97.5 to 104.2 percent of label as a generic drug and less than 2 percent to 110 percent as a supplement.
How to read a manufacturing claim on a label
Four checks, each of which takes a minute.
- Is the claim just the legal minimum restated? Complying with 21 CFR Part 111 is required of every supplement maker in the United States. A badge announcing good manufacturing practice compliance announces the floor.
- Does it name an organization and a standard? Manufactured in a GMP facility names neither. A claim worth something identifies who checked and against what.
- Is it about the facility or the product? Facility registration and product certification are different things, and the companion article on certifiers explains how to tell them apart.
- Can you check the firm’s record? FDA publishes its warning letters online, searchable by company name and date, with the full text of what investigators found. It is the one part of this record a member of the public can read without filing anything.
If a product you have taken is followed by a new symptom, the useful move is a clinical one rather than a regulatory one. Bring the actual bottle to the appointment, because the lot number on it is what connects a person to a batch, and a batch to a facility. Liver symptoms after starting a new supplement, which can include yellowing skin or eyes, dark urine, persistent nausea or pain under the right ribs, are the presentation that should not wait.
The most useful thing to know
The supplement manufacturing rule is real, detailed and enforceable, and the most common way to break it is to never have decided what your product is. That is not a technicality. A specification is the thing every other control in a factory is checked against, and roughly a third of cited firms in a typical year do not have one for their finished product.
The record that documents this is thin by design rather than by accident. FDA inspects a small share of facilities each year, publishes no supplement-specific inspection series, and relies on a letter as its most common enforcement tool. When that letter is sent, the published follow-up suggests the product usually stays on the shelf.
Keep exploring: supplement stacking and supplements and heart disease.
Sources
- Cornell Law School, Legal Information Institute. “21 CFR 111.70: What specifications must you establish?.” law.cornell.edu
- Cornell Law School, Legal Information Institute. “21 CFR Part 111: Current Good Manufacturing Practice in Manufacturing, Packaging, Labeling, or Holding Operations for Dietary Supplements (subparts A through P).” law.cornell.edu
- U.S. Food and Drug Administration. “Backgrounder on the Final Rule for Current Good Manufacturing Practices (CGMPs) for Dietary Supplements.” fda.gov
- U.S. Food and Drug Administration. “Inspection Observations (Form FDA 483 observation data, FY2006-FY2025, with agency caveats).” fda.gov
- U.S. Food and Drug Administration. “Using Dietary Supplements.” fda.gov
- U.S. Food and Drug Administration. “Warning Letter: Formulation Technology, Inc., 9 July 2024 (ref 672905).” fda.gov
- U.S. Food and Drug Administration. “Warning Letter: Professional Botanicals, Inc., 6 July 2017 (ref 517911).” fda.gov
- JAMA (PubMed). “Cohen PA, Avula B, Katragunta K, Khan I. "Recalls, Availability, and Content of Dietary Supplements Following FDA Warning Letters." 2022;328(4):393-395. DOI: 10.1001/jama.2022.9734.” doi.org
- JAMA Network Open (PubMed). “Cohen PA, Avula B, Katragunta K, Travis JC, Khan I. "Presence and Quantity of Botanical Ingredients With Purported Performance-Enhancing Properties in Sports Supplements." 2023;6(7):e2323879. DOI: 10.1001/jamanetworkopen.2023.23879.” doi.org
- SupplySide Supplement Journal (reporting FDA data obtained under the Freedom of Information Act). “Top FDA cGMP citation in FY23 tied to product specifications (8 January 2024).” supplysidesj.com
- SupplySide Supplement Journal (reporting FDA data obtained under the Freedom of Information Act). “FDA data: supplement cGMP compliance rates show modest improvement (27 February 2014), carrying the FY2012 and revised FY2013 rates.” supplysidesj.com
- SupplySide Supplement Journal (reporting FDA data obtained under the Freedom of Information Act). “FDA still finding same cGMP deficiencies at dietary supplement facilities (1 November 2018), carrying the FY2017 and FY2018 figures.” supplysidesj.com
- SupplySide Supplement Journal (reporting FDA data obtained under the Freedom of Information Act). “FDA supplement inspections highlight common quality control problems (4 January 2022), carrying the FY2021 figures.” supplysidesj.com
- SupplySide Supplement Journal. “FDA audits climb in FY22 for dietary supplement facilities (12 October 2022), carrying Cara Welch's statements on registered facilities and inspection coverage.” supplysidesj.com
- ConsumerLab.com (reporting FDA data obtained under the Freedom of Information Act). “FDA Supplement Manufacturing cGMP Inspections (10 September 2026), carrying the FY2025 figures.” consumerlab.com
- Nutritional Outlook. “FDA Holds Dietary Supplement Distributor, Not Contract Manufacturers, Ultimately Responsible for GMP (quoting FDA's 2011 warning letter to Confidence Inc.).” nutritionaloutlook.com
- Consumer Healthcare Products Association (hosting a presentation by EAS Consulting Group). “Summary of Key Differences Between FDA Drug and Dietary Supplement GMPs.” CHPA.org