How to Read a THCA COA Before Trusting the Potency

Pass one: prove it is the right report

I start at the top of the certificate, not the biggest cannabinoid number. I compare the product name, batch or lot identifier, sample ID, and any package identifier with the item being sold. If I cannot connect the report to the product, I stop treating it as product-specific evidence.

This is the most important lesson from my Golden Gardens Blue Dream listing audit: a public COA library is useful, but a visible batch match is what closes the loop. My Black Tie CBD review shows the stronger version with a full report linked on the product page. It also explains why a batch-number field marked “N/A” still leaves room for improvement.

Pass two: build the date trail

I note when the sample was received, when testing began, when the report was issued, and, if available, when the product was harvested or packaged. Those dates answer different questions. A current PDF upload does not make an old sample new, and an expiration date does not prove the seller still has the same batch.

For flower, I want to know whether the COA plausibly belongs to the inventory available today. If the seller rotates batches, the product page or package should make that relationship easy to verify.

Pass three: read the table headings and units

Cannabinoid results may appear as percent by weight, milligrams per gram, milligrams per unit, or another basis. I never compare two values until the units match. I also look for LOD and LOQ: the limits of detection and quantitation. “ND” generally means not detected above the method’s threshold; it does not mean the universe contains exactly zero molecules of that analyte.

For THCA flower, I look for THCA, delta-9 THC, and the lab’s total-THC calculation. The common estimate is delta-9 THC + (0.877 × THCA). My THCA chemistry explainer explains why that factor appears.

Pass four: identify what was actually tested

A potency panel is not automatically a safety panel. I scan the report index for the analyses that fit the product type: pesticides, heavy metals, microbial impurities, mycotoxins, foreign material, moisture or water activity, and residual solvents where relevant.

Testing requirements differ by jurisdiction and product category. California’s licensed-cannabis program, for example, lists cannabinoids, residual solvents, pesticides, heavy metals, microbial impurities, mycotoxins, moisture or water activity, and foreign material among required testing categories. That list is a useful illustration, not a universal checklist for every hemp product sold nationally.

I also read the result columns instead of accepting a storefront summary that says “full panel.” A report can include a section without testing the same analytes, limits, or methods used somewhere else.

Pass five: inspect the laboratory and method

I look for the laboratory name and address, signatures or authorization, accreditation information, analytical method, and report-page count. USDA strongly encourages ISO/IEC 17025 alignment for hemp testing and requires quality-assurance procedures for production compliance labs. California requires its licensed cannabis laboratories to maintain ISO/IEC 17025 accreditation.

Accreditation improves confidence in competence and process; it does not make every result infallible. NIST’s interlaboratory work has documented meaningful measurement variability and method-related bias in cannabinoid testing. That is one reason I treat a 29% result as a measurement with context, not a sacred ranking.

The red flags I write down

I flag a missing batch number, a report for a different product form, a certificate much older than the listed inventory, cropped pages, unreadable limits, a potency-only report presented as comprehensive safety testing, and any mismatch between the product-page headline and the certificate.

I also flag overconfident language. A passing contaminant panel means the tested sample met the listed limits for the listed analyses. It does not establish medical safety, predict effects, or settle every legal question.

My two-minute version

If time is short, I make five checks: same batch, sensible dates, matching units, visible test scope, identifiable laboratory. If any one of those fails, I slow down rather than letting the highest THCA number make the decision.

Next, read why a high THCA percentage does not guarantee better flower. It explains what NIST and state regulators have found about measurement variability and “lab shopping.”