You’re on a product page. The description says “third-party lab tested,” there’s a link to a report, and you click it. What opens is a PDF with a lab logo, a table of abbreviations, a column of numbers, and the word PASS in green. It looks reassuring. It also doesn’t explain itself.
This guide walks through that document from the top down: who issued it, what each section checks for, how to read the numbers, and what the report can’t tell you. You don’t need a lab background. Ten minutes and a report to look at will do, and the Lab Results page has plenty to practice on.
Table of contents
- What a certificate of analysis actually is
- Step 1: Check the header
- Step 2: Understand the test panels
- Step 3: Make sense of the numbers
- Step 4: What “potency” means for different plants
- What third-party lab results don’t prove
- Red flags worth noticing
- Putting it together: a two-minute COA check
- Frequently asked questions
- One report at a time
What a certificate of analysis actually is
A certificate of analysis, usually shortened to COA, is a document from a laboratory stating what it found when it tested one specific batch of a product. Think of it as a receipt from the lab. It records what was submitted, which tests were run, and what came back.
The “third-party” part matters. It means the lab is independent from the seller and from the supplier. Nobody with a stake in the sale ran the test. That independence is the reason the document carries weight.
In-house testing isn’t worthless. Plenty of manufacturers run their own checks as part of normal quality control. But a report produced by the company selling you the product is a different kind of evidence than one produced by an outside lab with its own accreditation to protect.
One more thing to set straight early. A COA is a snapshot of one batch on one date. It is not a standing guarantee about every bag of the same product.
“Lab tested” vs. “third-party lab tested”
These phrases get used interchangeably, and they shouldn’t be. “Lab tested” can mean tested anywhere, including the seller’s own facility. “Third-party lab tested” means an outside lab did the work. The quickest way to tell which one you’re looking at is to find the lab’s name on the document. If there’s no lab name, you don’t have a third-party report, whatever the label says.
Step 1: Check the header
The top of a COA answers three questions before you reach any results: who tested this, what exactly did they test, and when.
Lab name and accreditation
Look for the lab’s name, its address, and an accreditation mark. The one you’ll see most often is ISO/IEC 17025, an international standard for testing laboratories. In plain terms, accreditation means an outside body has audited the lab’s methods, equipment, and record-keeping and found them competent.
That raises confidence. It doesn’t make a lab infallible. Accredited labs can still make mistakes, and accreditation covers specific methods rather than everything a lab might do. Still, a report from an accredited lab is a meaningful step up from one without.
Batch or lot number
If you only check one thing, check this. The batch number (sometimes “lot number”) is the product’s name tag. It ties the report to a specific production run, and the same number should appear on the physical product’s label. If the numbers don’t match, the COA you’re reading isn’t for the item in your hand, no matter how clean it looks.
Purely Found product pages attach COAs by batch, so you can compare directly. Open any product in the kava collection alongside a bag of the same kava, and the numbers should line up.
Dates and sample description
A COA lists when the sample was received and when it was tested. Older reports aren’t automatically wrong, but a report from two years ago says nothing about stock produced last month. Look for a date that makes sense for what’s currently being sold.
The sample description tells you what form the lab received: raw powder, dried whole material, an extract, a finished capsule. This matters more than people expect. Testing a raw ingredient before it was processed doesn’t tell you much about what happened during processing.
Step 2: Understand the test panels
A COA is built out of panels. Each panel answers a different question, and a lab only runs the panels it was asked to run. The common ones, and why they exist:
Potency and identity
Is this the plant it claims to be, and roughly how much of the compounds of interest does it contain? Identity testing confirms the species. Potency testing measures specific compounds. What “potency” means changes from plant to plant, which is why it gets its own section below.
Heavy metals
Lead, arsenic, cadmium, and mercury are the usual four. Plants pull minerals from soil, and some soils carry more of these than others; the FDA has a plain overview of how metals end up in food and plant products. Results are typically reported in parts per million against a limit. This panel is relevant for almost any dried plant material.
Microbial contaminants
Total aerobic count, yeast and mold, E. coli, Salmonella. Dried botanicals and mushrooms can hold moisture, and moisture is where microbes grow. A microbial panel is one of the most important checks for anything sold as a tea, powder, or raw dried material.
Pesticides
A screen for residues from cultivation. Worth knowing: pesticide panels vary a lot in scope. One lab might screen for 60 compounds, another for several hundred. The COA should list which ones were included.
Mycotoxins
Aflatoxins and related compounds are produced by certain molds on stored plant material. This panel matters for dried botanicals and mushroom products in particular, since both can sit in storage for a while before sale.
Residual solvents
This one is only relevant for extracts. If a product was made by pulling compounds out of plant material with a solvent, the COA should show how much of that solvent is left. A tea or raw powder may reasonably not have this panel. An extract without it is worth a question.
A note on scope before moving on. A clean COA with two panels is not the same as a clean COA with six. “Pass” only means the product passed the tests that were run.
Step 3: Make sense of the numbers
This is where most people close the PDF. The abbreviations are fewer than they look.
| Term | What it means |
|---|---|
| ppm | Parts per million. 1 ppm is about 1 milligram per kilogram. |
| mg/g | Milligrams per gram. Often used for potency results. |
| % | Percent by weight. Also common for potency. |
| ND | Not detected. Below what the method can reliably find. Not the same as zero. |
| LOD | Limit of detection. The smallest amount the method can detect at all. |
| LOQ | Limit of quantitation. The smallest amount the method can measure with confidence. |
| Action limit | The threshold a result is compared against to decide pass or fail. |
Units: ppm, mg/g, and percent
Contaminant results (heavy metals, pesticides, solvents) usually show up in ppm or ppb. Potency results usually show up in mg/g or percent. Microbial results use colony-forming units per gram, written CFU/g. Different panels, different units. The COA should state which is which.
“ND,” LOQ, and LOD
“Not detected” is the result people misread most often. It means the lab’s method couldn’t find the compound above its detection limit. It does not mean there’s a true zero in the sample. Every method has a floor, and ND means the result is somewhere below that floor. That’s usually a good result. It just isn’t the same as “none.”
LOD and LOQ describe where that floor is. A thorough COA lists both so you can see how sensitive the test was.
Action limits and “pass”
A result passes when it falls below a limit. Those limits aren’t universal. Some come from regulatory guidance for dietary supplements, some from state programs, some from a brand’s own specifications. The FDA’s current good manufacturing practice rules set expectations for supplement manufacturers but don’t publish a single contaminant limit for every botanical. If a limit on a COA seems unusually high or low, asking what standard it’s based on is a fair question.
Step 4: What “potency” means for different plants
Potency numbers describe what’s in the batch. They don’t describe what you’ll feel, and they don’t suggest how much to take. Two people can have very different experiences with the same number on the same report. Keep that in mind through this section.
Kava
Kava COAs commonly report total kavalactones as a percentage, and some break down the six major kavalactones into a chemotype. Kava testing is relatively well established compared to the other plants here, with methods that have been in use for years. That makes kava a good first COA to practice on. The kava collection is a reasonable place to find one.
Amanita Muscaria
Reports for Amanita Muscaria may list muscimol and ibotenic acid, the two compounds most often measured. Testing for Amanita is less standardized than testing for kava, and that’s worth saying plainly. Reference standards are newer, methods differ between labs, and results can shift depending on how the mushroom was dried or prepared. Research on this mushroom is still developing, and the testing is developing alongside it. Treat Amanita potency numbers as useful information from one lab on one batch, not as a settled figure.
Kanna
Kanna COAs may show total mesembrine-type alkaloids. These alkaloid profiles vary with cultivar and preparation, so two kanna products can look quite different on paper. Kanna (Sceletium tortuosum) has a long history of traditional use among the Khoikhoi and San peoples of southern Africa.
Blue lotus
Blue lotus testing may report nuciferine and related aporphine alkaloids, but many blue lotus products are tested primarily for contaminants rather than potency. That’s an honest reflection of where the testing is right now. A contaminant-only COA for blue lotus isn’t a red flag on its own.
Mushroom blends
A quick clarification, since the term causes confusion. “Mushroom blends” at Purely Found means species like lion’s mane, reishi, chaga, and cordyceps, sometimes called functional mushrooms. These are culinary and traditional-use species. They are not Amanita, and they are not any kind of psychoactive mushroom.
COAs for these products often report beta-glucan percentage, a class of compounds the mushrooms are commonly standardized to. Pay attention to whether the material is fruiting body (the mushroom itself) or mycelium grown on grain. Mycelium-on-grain products can carry a lot of leftover grain, which affects what the numbers represent. A COA that specifies the material type is giving you more to work with.
What third-party lab results don’t prove
A COA is a good tool with real limits. Here is what it can’t do:
- Predict an experience. Potency numbers describe the batch, not what any particular person will feel.
- Confirm a product is right for you. Anyone who is pregnant, nursing, taking medication, or managing a health condition should talk with a qualified healthcare professional before trying any botanical. A clean COA doesn’t change that.
- Tell you how much to take. Nothing on a lab report is a serving suggestion.
- Make a health claim. A COA measures what’s in a batch. It isn’t evidence that a product can diagnose, treat, cure, or prevent anything, and nothing in this guide should be read that way.
- Speak for other batches. The report covers the tested batch and the tested panels. That’s it.
- Tell you what’s legal where you live. Laws around botanical and plant-based products vary by state and locality, and it’s on the buyer to know the rules in their area before ordering.
- Replace the basics. Reading the label, starting gently, and paying attention to your own experience still matter.
The Disclaimer page covers this in full.
Red flags worth noticing
Most COAs are exactly what they appear to be. A few things should make you pause:
- No lab name, address, or accreditation anywhere on the document
- No batch number, or a batch number that doesn’t match the product label
- Results dated well over a year ago with no newer batch available
- Panels missing that the product type calls for (an extract with no solvent panel, for example)
- A report that looks typed into a word processor rather than issued by a lab, with no signature or report ID
When something doesn’t add up, ask the seller. A seller who can’t answer, or won’t, is its own data point.
Putting it together: a two-minute COA check
- Match the batch number on the report to the one on the product.
- Find the lab’s name and accreditation.
- Check the test date against what’s currently being sold.
- Scan the panels and make sure the ones this product type needs are there.
- Read the results against their limits, remembering that ND isn’t zero.
- Remember what the report doesn’t tell you.
Frequently asked questions
"Lab tested" could mean tested anywhere, including by the seller. "Third-party lab tested" means an independent lab ran the analysis. Look for the lab's name on the report.
No. "Not detected" means the amount is below what the test method can reliably find. It's usually a good result, but it isn't a true zero.
Ideally every batch. At minimum, you should be able to find a report that matches the batch number on the product you're buying.
Yes. Reach out through the Contact page with the batch number and we'll send the matching report.
Lab results address contaminants and content in a specific batch. They don't address personal fit. If you're pregnant, nursing, on medication, or managing a health condition, a conversation with a healthcare professional is the right next step.
One report at a time
Reading third-party lab results is a skill you pick up, and it comes quickly. The first one takes ten minutes. The second takes five. After that, you’ll find yourself glancing at the batch number and the panel list out of habit.
Open any report on the Lab Results page and try it. And if anything on a report doesn’t make sense, ask us about the batch. We’d rather explain a number than have you guess at it.