How to Read a Certificate of Analysis (COA)

What a COA can—and cannot—tell you

A Certificate of Analysis, or COA, is a laboratory report describing what was measured in a particular sample. For a hemp product, it may show cannabinoid concentrations and results for contaminants such as pesticides, heavy metals, residual solvents, and microorganisms.

A useful COA connects those measurements to the product in your hands. It helps you compare a label’s claims with laboratory findings and understand the limits of the testing.

It is not a government approval, a guarantee about every bottle, or evidence of a wellness outcome. The most informative question is not simply “Does this product have a COA?” It is “What was tested, how was it measured, and does this report match my product?”

Start by matching the report to the product

Before reading the cannabinoid table, check the identifying information.

Look for:

• Product name and form: oil, gummy, capsule, or another finished product.
• Batch or lot number: ideally matching the number on your packaging.
• Sample identification: the laboratory’s reference for the submitted material.
• Relevant dates: when the sample was received, tested, and reported.
• Laboratory identity: its name, contact information, and report number.

A report for hemp extract is not necessarily a report for the finished gummy or bottled oil made from it. Formulation and manufacturing introduce additional ingredients and steps. Finished-product testing therefore provides a more direct view of what you are buying.

If a QR code leads to a generic report with no matching batch number, ask the brand for the correct document.

### Read dates in context

A recently posted PDF is not necessarily a recently tested sample. Check the dates inside the report.

Older results may still correspond to a product within its stated shelf life, but they do not automatically demonstrate stability throughout storage. Heat, light, packaging, and time can affect product composition. Stability testing is a separate question from a single batch’s release testing.

Check the laboratory’s credentials

An independent laboratory provides separation between the manufacturer and the measurement process. Independence alone, however, does not establish technical competence.

Look for accreditation to ISO/IEC 17025, the international standard for testing and calibration laboratories. ISO describes this standard as addressing laboratory competence, impartiality, and consistent operation.

Accreditation has a defined scope. A laboratory may be accredited for certain methods, substances, or sample types—not every test it offers. When available, check the laboratory’s accreditation certificate and scope through the accrediting body.

Also look for named analytical methods. Cannabinoid testing commonly uses liquid chromatography, while other instruments and methods are used for metals, solvents, and microorganisms. You do not need to become an analytical chemist; you need enough information to see that the result comes from a documented process.

Decode the cannabinoid table

The cannabinoid profile lists compounds the laboratory tested for, which may include CBD, CBDA, delta-9 THC, THCA, CBG, and CBC.

Not every product will contain measurable amounts of every listed cannabinoid. The table is a testing panel, not an ingredient list.

Pay close attention to units:

• mg/g: milligrams per gram of sample.
• mg/mL: milligrams per milliliter.
• % by weight: grams per 100 grams of sample.
• mg/unit or mg/serving: amounts calculated for a defined product unit or serving.

For results expressed by weight, 1% equals 10 mg/g. But converting mg/g to mg/mL requires the product’s density. Do not assume one gram of oil equals one milliliter.

### Compare like with like

Imagine a report lists 25 mg of CBD per gummy. If the package contains 20 gummies, that corresponds to 500 mg per package, assuming the tested concentration represents those units.

Now imagine the report instead lists 25 mg/g. You would need the gummy’s weight to calculate CBD per gummy.

Check whether the laboratory measured the finished product or whether a serving value was calculated from another measurement. Small differences also require context: sampling, rounding, and measurement uncertainty can influence results.

Understand “ND,” detection limits, and quantitation limits

A result marked ND usually means “not detected.” It does not prove that absolutely none of the substance exists.

Two terms help explain why:

• Limit of detection, or LOD: the lowest level a method can reliably distinguish from background.
• Limit of quantitation, or LOQ: the lowest level a method can reliably quantify with acceptable performance.

A substance may be detectable but below the level that can be quantified reliably. Laboratories may report this as “<LOQ,” “detected below LOQ,” or another notation.

Always read the report’s legend. Reporting conventions vary, and an ND result without an accompanying reporting limit gives you less information than it appears to.

Read THC results with particular care

Extract terminology can provide context, but it cannot replace testing.

Broad-spectrum extracts generally contain multiple hemp constituents with THC removed or reduced. Full-spectrum extracts generally retain THC alongside other constituents. Isolates focus on an individual cannabinoid. These terms alone do not establish a sample’s measured composition.

Divine Earth’s products are broad-spectrum and labeled 0.0% THC. When evaluating that claim, examine the relevant THC entries and the laboratory’s reporting limits. A percentage rounded to one decimal place is not the same as proof of absolute molecular absence.

Some reports also list total THC, a calculated value accounting for delta-9 THC and the potential contribution of THCA after conversion. Read the calculation notes rather than treating delta-9 THC and total THC as interchangeable.

A COA cannot guarantee a negative workplace drug test. Nor does a THC result by itself settle every legal question. As of 2026, hemp regulation remains subject to change; check current state law and applicable federal requirements.

Look beyond potency to contaminant testing

A cannabinoid-only report answers a composition question. It does not establish that the product was tested for contaminants.

Look for separate panels covering relevant categories:

• Pesticides: the specific agricultural chemicals included in the panel.
• Heavy metals: often including lead, cadmium, arsenic, and mercury.
• Residual solvents: chemicals potentially remaining from extraction or processing.
• Microbial contaminants: specified organisms or microbial counts.
• Mycotoxins: certain toxins produced by molds.

Testing needs can differ by ingredients, manufacturing process, and product type. An oil and a water-containing edible do not necessarily present identical testing considerations.

### Ask what “pass” actually means

“Pass” means the reported result met the criterion applied in that report. Find the action limit or specification and, where provided, its regulatory or technical basis.

“Not tested” is not equivalent to “not detected.” A report showing acceptable results for several metals says nothing about pesticides if no pesticide panel was performed.

Even comprehensive testing covers a defined list of substances. It is not a search for every possible contaminant.

Recognize the limits of the evidence

Laboratory quality matters because cannabinoid measurement is technically demanding. The National Institute of Standards and Technology’s Cannabis Quality Assurance Program helps laboratories evaluate and improve measurement comparability.

Research also supports checking labels against measurements. A study published in JAMA in 2017 documented labeling inaccuracies among CBD extracts sold online. That study does not establish the accuracy of any current brand, but it illustrates why independent verification matters.

The FDA has likewise raised concerns about some CBD products’ labeling and potential contaminants. A COA does not mean a product is FDA-approved.

Finally, remember that the laboratory tested a submitted sample. Unless sampling procedures are documented, the report may not show how that sample was selected or how representative it was of the batch.

The takeaway

A strong COA is traceable, readable, and specific: it matches your batch, identifies the laboratory, explains its units and reporting limits, and includes relevant contaminant testing.

Read beyond the bold “pass” and the CBD number. The value of a COA lies in both what it demonstrates and how clearly it defines what remains unknown.