You buy the same supplement twice, a year apart. Same brand, same label, same number printed on the front. Most of us assume those two bottles are the same thing.
They might not be close.
Researchers at the University of Guelph in Ontario analyzed 31 commercial melatonin supplements and published the results in the Journal of Clinical Sleep Medicine in 2017. Melatonin content ranged from 83% below the labeled amount to 478% above it. More than 71% of products missed their own label claim by more than 10%. And here’s the number that reorganizes how you should think about all of this: within a single product, from one production lot to another, melatonin content varied by as much as 465%.
Same product. Same label. Different bottle, different dose.
That finding isn’t about one hormone or one company. It’s about a fact of manufacturing that almost no shopper has been told. A supplement isn’t a single object with fixed properties. It’s a series of production runs, and a lab test describes the run it was performed on. Understanding that changes what a test result is worth to you.
A test is a question about one sample, not a product
Every laboratory test has three parts: a sample, a method and a specification.
The sample is the physical material someone put in the instrument. The method is the procedure used to measure it. The specification is the number the result gets compared against, meaning the limit the manufacturer wrote down in advance.
Change any one of those three and the answer can change. That’s not a scandal, it’s measurement. But it means a sentence like “this product was tested” is incomplete to the point of being unusable. Tested which lot? By what method? Against whose specification? What exactly the phrase “third-party tested” establishes is covered in its own article in this library. This one is about what happens in the lab.
The main kinds of test and what each one answers
Identity: is this the ingredient?
Identity testing answers the first and most basic question. Federal rules require it. Under 21 CFR 111.75, a manufacturer has to conduct at least one appropriate test to verify the identity of any dietary ingredient before using it, unless FDA grants an exemption through a petition.
Methods range from simple to sophisticated: visual and microscopic examination, chemical analysis, chromatography, and for botanicals, DNA sequencing. A 2020 study in PLOS ONE used DNA barcoding to authenticate seven reishi mushroom supplements from independent suppliers and found the expected species in all of them. A 2016 study in Food Additives and Contaminants ran the same kind of analysis on five herbal products sold in the Philippines and found only one that met both of the authentication criteria the researchers set.
Those are small studies, and small studies are weak evidence about a market. What they do show is that identity is a real, answerable scientific question, which matters because a botanical powder gives up very few clues to the eye.
Potency: how much is in there?
Potency testing, often called assay testing, measures how much of a marker compound a serving contains. This is the number buyers care about most, and it’s the one most dependent on method choice. For a plant extract there’s often no single compound that defines potency, so the lab measures markers and the answer depends on which markers were chosen.
Purity: what else is in there?
Contaminant screening looks for heavy metals, microbes, pesticide residues and undeclared drug ingredients. That subject is large enough to have its own article here, so take this as the testing-side note: screening finds what the method was designed to look for and nothing else. A clean heavy-metals panel says nothing about a drug ingredient nobody tested for.
Disintegration and dissolution: does the pill break down?
A capsule that doesn’t open delivers nothing. Disintegration testing asks whether the dosage form breaks apart. Dissolution testing asks whether the compound goes into solution.
This is where batch variation shows up in the clearest way anyone has published. Scientists at USDA’s Agricultural Research Service and NIH’s Office of Dietary Supplements tested 28 single-ingredient green tea supplements, two to four lots of each, and reported in the Journal of Pharmaceutical Sciences in 2020 that 9 products, about 32%, always passed disintegration; 8 products, about 29%, always failed; and 11 products, about 39%, gave inconsistent results from lot to lot. For dissolution, only 6 of 20 single-ingredient products released enough of the main green tea compound to pass.
The authors added a caution that deserves equal weight: the general pharmacopeial protocols may not be well suited to this particular botanical. A failed test can mean the product is bad. It can also mean the method is wrong for the material. Both happen.
Stability: how much is in there later?
Stability testing asks what the product contains at the end of its shelf life rather than at the moment it was made. It’s the test most often skipped, and it’s the reason a result from manufacture doesn’t automatically describe a bottle that spent 14 months on a warm shelf.
Batch-to-batch variation, the idea most shoppers are missing
A batch, also called a lot, is one production run. Federal rules treat it as the basic unit of a supplement’s existence.
Under 21 CFR 111.255 and 111.260, a manufacturer has to create a batch production record for every single batch. That record carries the batch, lot or control number, the identity and weight of every component, the equipment used, the actual yield compared to the expected yield, the test results, and the quality control decision to release or reject. The government’s framework assumes from the start that each run is its own event with its own paperwork.
Lots differ for ordinary reasons. Botanical raw material changes with growing region, weather, harvest timing and which part of the plant was used. Powders separate and blend unevenly. A supplier changes. A machine is recalibrated. An overage that compensated for expected degradation gets adjusted.
The consequence for you is specific. A certificate of analysis is a document about one lot. A magazine’s test result is about one lot. A certifier’s listing is about the products and lots that program covers. None of them is a statement about the bottle in your hand unless the lot numbers match.
Look at your bottle now. There’s a lot or batch number on it, usually stamped near the expiration date, often in faint ink. That little string is the only thing connecting any test result to the product you own. If a company publishes certificates of analysis, you can ask for the one matching your lot. If it only offers a generic certificate with no lot number, that tells you something.
Why two labs get two numbers from the same sample
Send one sample to two good laboratories and you’ll often get two different numbers. That’s expected, and the size of the gap is larger than most people would guess.
The clearest demonstration comes from collaborative studies, where many labs analyze identical blind samples using the same validated method. One such study, published in the Journal of AOAC International in 2009, had 12 laboratories measure three Aconitum alkaloids, toxic compounds that can appear in certain botanical materials, across 14 blind duplicate samples. Within a single lab, results varied by roughly 2% to 17%. Between labs, the spread ran from about 7% to 33%.
Read those two ranges together. Every lab was competent, every lab used the same procedure, and between-lab variation was still several times the within-lab variation at the high end. That’s not sloppiness. It’s what measurement looks like when you examine it closely, and it’s the same reason your cholesterol value isn’t directly comparable across two different labs, a point covered in this library’s article on reference ranges.
Laboratories manage this problem with reference materials: physical samples whose contents have been characterized by multiple methods, used to check that an instrument and a procedure are producing truthful numbers. NIH’s Office of Dietary Supplements runs a program that has helped develop more than 60 of them, mostly through work with the National Institute of Standards and Technology, covering things like ginger rhizome, Asian ginseng extract and chromium supplements.
The program is candid about the limits. Those certified materials aren’t meant to serve as a gold standard or to set an industry specification. They’re built to reflect the messy reality of real samples, which is exactly why they’re useful and exactly why “the lab said so” is not a complete answer.
Why testing one bottle proves so little
Imagine testing one apple to describe an orchard. That’s roughly what a single-bottle test does.
Federal rules take sampling seriously for this reason. Under 21 CFR 111.80, manufacturers must collect representative samples at multiple stages, including components, in-process material and finished batches. For finished batches they can either test every batch or use a scientifically sound statistical sampling plan that covers a subset.
There’s also a mechanism almost nobody outside the industry knows about. Under 21 CFR 111.83, manufacturers must keep reserve samples of every lot they distribute, stored in the same kind of container the product was sold in, held for one year past the shelf life date or for two years after the last batch was distributed. The physical material is supposed to still exist, so that if a problem surfaces later, someone can go back and test the actual lot.
That rule contains a quiet lesson for shoppers. The requirement is written as one year past the shelf life date “if shelf life dating is used.” Expiration dating on a supplement is not automatic, and a product without one gives you no way to judge whether stability was ever considered.
Manufacturing is the layer underneath the test
Testing only means something on top of consistent manufacturing. That’s the part of this subject that never appears in advertising.
FDA’s compliance guide for the dietary supplement manufacturing rule defines the goal in one line. Quality means the product “consistently meets the established specifications for identity, purity, strength, and composition” and stays within limits on contamination. Note the first word. Not meets once. Consistently.
The rule behind that guide, 21 CFR Part 111, was published in the Federal Register on June 25, 2007 and took effect that August, with compliance phased in by company size: the largest companies by June 2008, mid-sized ones by June 2009 and companies with fewer than 20 employees by June 2010.
What it requires is more detailed than most people expect. A master manufacturing record for every unique formulation and batch size, specifying the weight of each ingredient and the expected yield at each control point, with maximum and minimum limits that trigger an investigation when a run falls outside them. Two-person verification when an ingredient is weighed by hand. Quality control staff whose duties are separate from their other work and who must approve or reject processes and specifications. Incoming components quarantined until quality control releases them.
The rule even builds in skepticism toward paperwork. A manufacturer may rely on a supplier’s certificate of analysis for certain components instead of testing them, but only after qualifying that supplier and periodically re-confirming the certificate. Inside the federal rule itself, a certificate is a document you verify, not a document you trust.
The NIH Office of Dietary Supplements notes that FDA periodically inspects manufacturing facilities. But remember the structural fact underneath all of it: supplements don’t require FDA approval before sale, so the agency is generally reviewing a system after products are already on shelves.
Where you can see test data
Your realistic options are narrower than the marketing suggests.
Manufacturers may provide a certificate of analysis on request, and the one worth having is the one matching your lot number. In the Journal of the American Pharmacists Association in 2016, an international group of probiotic researchers argued that pharmacists stocking products for at-risk patients should require certificates of analysis documenting quality standards, and that third-party verification adds assurance about manufacturing and label accuracy. If that’s the standard proposed for hospital pharmacies, asking for it as an individual is not unreasonable.
Certifier registries let you confirm that a specific product currently holds a mark, which is the subject of a companion article here. Published analyses in the scientific literature are free to read in abstract form and often tell you more than any brand page, though they’re usually small, sometimes years old and rarely about the exact product you’re holding.
When you do find a result, four questions tell you what it’s worth:
- Which lot was tested? If the document doesn’t name one, it’s marketing with a chart.
- Who ran it, and against what specification? A number with no limit attached can’t pass or fail anything.
- What method was used, and is it appropriate for this material? The green tea study is the reminder that a method can fail a product unfairly.
- How old is the result compared to the bottle I own? A certificate from three years ago describes a production run that no longer exists.
One limit applies to every answer. A lab can confirm what’s in a capsule. It can’t confirm that what’s in it helps you. That question belongs to clinical trials, which this library covers separately. A product can be perfectly made, accurately labeled, verified by a credible program, and still do nothing for the reason you bought it.
Safety sits outside the lab report too. Two interactions from the Office of Dietary Supplements make the point. Warfarin becomes less effective alongside vitamin K, and many medicines break down faster alongside St. John’s wort. Disclose supplements before surgery, the office advises. If you react badly to a product, FDA says to stop using it immediately and file a report through the federal Safety Reporting Portal. The agency lists hives or throat swelling, fainting, chest pain or an irregular heartbeat, abnormal bleeding, yellowing of the skin or eyes, severe ongoing digestive symptoms and stroke symptoms among serious adverse events. Those call for medical care first and paperwork second.
The most useful thing printed on your bottle
It’s the lot number.
Not the seal, not the claim, not the certificate posted on a website. The lot number is the only thing that ties a laboratory result to the physical product you own, because testing describes production runs and your bottle came from exactly one of them.
So write it down, along with the expiration date, when you start a new container. If a recall is announced, you’ll know in seconds whether it’s yours. If you ask a company for test data, you’ll be asking a question it can answer. And if you ever see a test result for a product you take, you’ll know the right first question to ask, which is whether that result has anything to do with your bottle at all.
Sources
- U.S. Government Publishing Office (eCFR). “21 CFR Part 111: Current Good Manufacturing Practice in Manufacturing, Packaging, Labeling, or Holding Operations for Dietary Supplements.” ecfr.gov
- U.S. Food and Drug Administration. “Small Entity Compliance Guide: Current Good Manufacturing Practice in Manufacturing, Packaging, Labeling, or Holding Operations for Dietary Supplements.” fda.gov
- NIH Office of Dietary Supplements. “Analytical Methods and Reference Materials Program: Reference Materials.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Analytical Methods and Reference Materials (AMRM) Program.” ods.od.nih.gov
- NIH Office of Dietary Supplements. “Dietary Supplements: What You Need to Know.” ods.od.nih.gov
- U.S. Food and Drug Administration. “Dietary Supplements.” fda.gov
- U.S. Food and Drug Administration. “How to Report a Problem with Dietary Supplements.” fda.gov
- PubMed. “Erland LAE, Saxena PK. "Melatonin Natural Health Products and Supplements: Presence of Serotonin and Significant Variability of Melatonin Content." Journal of Clinical Sleep Medicine, 2017. DOI: 10.5664/jcsm.6462.” doi.org
- PubMed. “Gusev PA, Andrews KW, Savarala S, et al. "Disintegration and Dissolution Testing of Green Tea Dietary Supplements: Application and Evaluation of United States Pharmacopeial Standards." Journal of Pharmaceutical Sciences, 2020. DOI: 10.1016/j.xphs.2020.02.005.” doi.org
- PubMed. “Wong SK. "Determination of Aconitum Alkaloids in Dietary Supplements and Raw Botanical Materials by Liquid Chromatography/UV Detection with Confirmation by Liquid Chromatography/Tandem Mass Spectrometry: Collaborative Study." Journal of AOAC International, 2009. PubMed. PMID: 19382567.” ncbi.nlm.nih.gov
- PubMed. “Sanders ME, Merenstein DJ, Ouwehand AC, et al. "Probiotic use in at-risk populations." Journal of the American Pharmacists Association, 2016. DOI: 10.1016/j.japh.2016.07.001.” doi.org
- PubMed. “Gunnels T, Creswell M, McFerrin J, Whittall JB. "The ITS region provides a reliable DNA barcode for identifying reishi/lingzhi (Ganoderma) from herbal supplements." PLOS ONE, 2020. DOI: 10.1371/journal.pone.0236774.” doi.org
- PubMed. “Olivar JEC, Alaba JPEP, Atienza JFM, et al. "Establishment of a standard reference material (SRM) herbal DNA barcode library of Vitex negundo L. (lagundi) for quality control measures." Food Additives & Contaminants Part A, 2016. DOI: 10.1080/19440049.2016.1166525.” doi.org