Cannabis Lab Testing: How to Read a Certificate of Analysis

Everything on cannabis packaging is marketing except the certificate of analysis. The COA is a laboratory document, and it is the only place you can check what a product actually contains. Learning to read one takes about ten minutes and is the single highest-value skill for a cannabis consumer.

What Testing Is Supposed to Cover

Regulated markets differ in the details, but a full panel generally covers five areas.

1. Potency (cannabinoid profile). How much THC, CBD and, increasingly, minor cannabinoids the product contains. 2. Pesticides. Screening against a defined list of prohibited agricultural chemicals. 3. Residual solvents. Relevant for extracts, checking for leftover butane, propane, ethanol and similar processing agents. 4. Heavy metals. Usually lead, arsenic, cadmium and mercury. Cannabis is a bioaccumulator and readily takes metals up from soil. 5. Microbial and mycotoxin contamination. Yeast, mould, aspergillus, E. coli, salmonella, and the aflatoxins some moulds produce.

Some jurisdictions add water activity, moisture content, foreign matter or terpene profiling. Terpene panels are commonly optional, which is worth remembering when a product is sold on its terpene story.

Reading the Potency Panel

This is where most confusion lives, because there are two different THC numbers and they are not interchangeable.

Raw cannabis contains almost no active THC. It contains THCA, the acidic precursor, which converts to THC when heated. A COA reports both, plus a calculated total.

The industry-standard formula is:

Total THC = THCA x 0.877 + delta-9 THC

The 0.877 factor is a mass correction. THCA loses a carboxyl group during decarboxylation, and that lost CO2 accounts for roughly 12.3 percent of the molecule's mass. The same logic gives Total CBD = CBDA x 0.877 + CBD.

Two practical consequences follow:

A flower COA showing 0.8 percent delta-9 THC and 24 percent THCA is not weak. Its total THC is about 21.8 percent.

If a label quotes a number noticeably higher than the formula gives, either a different calculation is being used or the label is being generous. Do the arithmetic yourself.

Potency is also reported per unit rather than per package on edibles, which is where accidental overconsumption usually begins. A 100 mg chocolate bar scored into ten squares is ten 10 mg doses.

Reading the Contaminant Panels

Contaminant results are usually reported as Pass/Fail against an action limit, sometimes with the measured value. Useful things to look for:

ND or below LOQ. "Not detected" or "below the limit of quantitation" means the analyte was not found at a measurable level. This is what you want on pesticides and heavy metals.

Pass with a number close to the limit. A pass is a pass, but a result sitting near the action limit on heavy metals or mycotoxins is worth noting, particularly if you consume the product daily.

Total aerobic and total yeast and mould counts. Elevated counts matter most for immunocompromised consumers and for anyone smoking rather than eating the product.

Solvent panels only matter for extracts. A flower COA with no residual solvent section is not incomplete.

The Red Flags That Actually Matter

A COA with no batch or lot number. Without one it cannot be tied to the product in your hand, which makes it decorative.

A stale date. Cannabinoids degrade and testing is batch-specific. A COA from two years ago tells you little about this jar.

Only a potency panel. Potency-only testing is far cheaper than a full panel. Its absence of contaminant sections is the finding.

Sections marked "not tested". This is not the same as passing.

No named laboratory, or a lab with no accreditation. Look for ISO/IEC 17025 accreditation, which covers testing competence.

A QR code that resolves to a marketing page rather than a document. The link should reach a report, not a brand site.

Suspiciously round potency numbers. Real analytical results rarely land on exactly 30.0 percent.

Why Reported Potency Is Often Wrong

Potency inflation is a documented problem in legal markets, and it has structural causes rather than purely dishonest ones.

Sampling. Results describe the sample submitted, not the batch. A sample taken from the best flower in a harvest legitimately tests higher than the batch average.

Lab shopping. When retail price tracks THC percentage, producers have an incentive to keep using whichever lab reports the highest numbers.

Method variation. Sample preparation, extraction efficiency and instrument calibration all move results. The same material can differ by several percentage points between labs.

The reasonable conclusion is not that COAs are worthless. It is that the contaminant panels, which nobody has an incentive to inflate, are the more trustworthy and arguably more important half of the document.

Checklist Before You Buy

Before you buy, confirm: the COA exists and is reachable; the batch number matches the product; the date is recent; a named accredited lab performed it; contaminant panels are present and passed; and the total THC figure is consistent with the formula.

If a seller cannot produce a COA at all, you are not evaluating a product, you are trusting a stranger. That is the actual difference between the regulated and unregulated market.

Disclaimer: This article is for educational purposes only and does not constitute medical advice. Always consult with a qualified healthcare professional before making any health-related decisions.