Cannabis potency testing is the laboratory analysis that measures how much THC, CBD, and related cannabinoids a product actually contains before it can be sold at a licensed dispensary. Every regulated batch has to go through it, the results drive the numbers printed on the label, and the quality of those numbers depends heavily on which lab did the work and which state oversees them.
What Labs Measure
The compounds that matter most are tetrahydrocannabinolic acid (THCA) and cannabidiolic acid (CBDA), the acidic forms that exist in raw plant material. Neither is psychoactive on its own. Heat from smoking, vaping, or cooking triggers decarboxylation, which converts THCA into Delta-9 THC and CBDA into CBD. A useful potency result has to account for both the acidic precursors and the already-active forms, because together they predict what a consumer will actually feel.
Most regulatory frameworks also require screening for secondary cannabinoids such as cannabigerol (CBG) and cannabinol (CBN). CBN tends to rise as THC degrades, so its level can indicate how fresh the product is. These minor compounds appear in smaller concentrations and rarely make it onto labels: only about 16% of products on the market list information about lesser-known cannabinoids beyond THC and CBD.
How the Test Is Run
High-Performance Liquid Chromatography (HPLC) is the standard method. The lab dissolves a prepared sample in a liquid solvent and pumps it through a pressurized column packed with adsorbent material. Different cannabinoids travel through the column at different speeds, and a detector measures how much light each one absorbs at specific wavelengths. Software turns those readings into concentration percentages.
The reason HPLC wins out over older methods is temperature. HPLC runs at room temperature, which means THCA stays as THCA and THC stays as THC through the entire analysis. Gas chromatography (GC), the alternative, vaporizes the sample through a heated injection port, and that heat forces THCA to convert into THC during the test itself. GC can also trigger side reactions like oxidation. Some labs still use GC and apply mathematical corrections to estimate original THCA, but that adds a layer of imprecision HPLC avoids.1Chromatography Online. Perspectives and Pitfalls in Potency Testing of Cannabinoids by High Performance Liquid Chromatography GC still has a place for residual solvent and contaminant work.
Before anything goes into the instrument, technicians grind the plant material to a uniform consistency. Cannabinoids are not evenly distributed across a bud; trichome-dense surfaces test far higher than stems and leaves, so homogenization matters. The ground material is then extracted into a solvent, filtered, and diluted to fit the detector’s working range. Sloppy preparation can throw off a result as badly as a miscalibrated machine.
From Raw Data to the Label
Raw instrument readings do not go straight onto packaging. Because most THC in unheated cannabis exists as THCA, labs apply a standard formula to predict the total THC a consumer will actually experience after decarboxylation:
Total THC = (THCA × 0.877) + Delta-9 THC
The 0.877 factor reflects the share of the THCA molecule’s weight that is lost as carbon dioxide when the acid converts. Not all THCA converts in real use, so the formula produces a theoretical maximum rather than a guaranteed delivery. The same calculation applies to CBD: Total CBD = (CBDA × 0.877) + CBD.
Each tested batch generates a Certificate of Analysis (COA). A COA lists the full cannabinoid profile, contaminant screening results for pesticides, heavy metals, microbial impurities, and residual solvents, and a pass or fail determination for each test. It also records the batch number, sample type, collection date, and the lab’s license credentials. Some states require dispensaries to make COAs accessible to consumers through QR codes on the packaging.
States set tolerance windows for how far a product’s measured potency can drift from what the label claims, typically ±10% to ±15% of the stated cannabinoid concentration. Low-dose products sometimes carry wider tolerances because small absolute measurement errors translate into large percentage swings at a 5 or 10 milligram target.
How Reliable Are the Numbers?
Labels are less dependable than the testing system suggests. A 2025 peer-reviewed study found that nearly half of cannabis flower products were inaccurately labeled for THC potency, and most of the inaccurate ones overstated strength. Concentrates performed much better, with about 96% matching their labels within tolerance. CBD labeling compliance was moderate, and information about minor cannabinoids was almost entirely absent from packaging.2PMC (PubMed Central). Commercial Cannabis Product Testing: Fidelity to Labels and Regulations
The driver is lab shopping. Flower testing below 20% THC struggles to move off dispensary shelves, so producers gravitate toward labs that return higher numbers. Industry estimates suggest roughly 30% of cannabis testing labs in the United States report inflated results, and because those labs attract more business, they end up testing a disproportionate share of the market. Flower labeled at 30% to 45% THC has shown up on shelves, concentrations plant biology makes essentially impossible. One analysis estimated that consumers in a single state paid for $90 million worth of THC that didn’t actually exist in their products in a single year. The pattern is visible in state data: THC distributions across tested products show an unnatural cliff right at 20%, with very few products landing just below, which is what you would expect if borderline results were being nudged upward.
States have started responding. California now requires all licensed testing labs to use the same standardized operating procedures and analytical methods. Oklahoma created a state reference laboratory and runs a proficiency testing program to audit results. Maryland is increasing audit frequency and plans to use a reference lab to retest products with suspiciously high THC values. Several states have revoked lab licenses over evidence of inflated reporting, and at least one group of labs has filed private litigation against competitors for producing fraudulent results.
The practical reading for a consumer: the THC percentage on a flower label is an estimate with real uncertainty behind it, and in some markets an optimistic one. Concentrate and edible labels tend to track reality more closely.
Measurement Uncertainty
Even an honest lab doesn’t produce a single perfect number. Every potency result carries a margin of error called measurement uncertainty (MU). If a lab reports 22% THC with an MU of ±1.5%, the true value likely sits between 20.5% and 23.5%.
Under ISO/IEC 17025, accredited labs must identify every significant source of uncertainty, evaluate it during instrument calibration, and estimate it for every result. For hemp testing under the federal domestic hemp production program, labs must report MU as a plus-or-minus value in the same unit as the regulatory threshold. The USDA does not fix a single formula for how MU should be calculated and instead relies on performance standards like the AOAC Standard Method Performance Requirements.3Agricultural Marketing Service. Laboratory Testing Guidelines U.S. Domestic Hemp Production Program
MU matters most at regulatory boundaries. A hemp sample testing at 0.31% Delta-9 THC may still be compliant if its uncertainty range reaches below 0.3%. Without a reported MU, a result that looks like a failure might fall within the legal window, and the opposite is also true.
What Makes a Testing Lab Legitimate
ISO/IEC 17025 Accreditation
Regulated states require cannabis testing labs to hold ISO/IEC 17025 accreditation, the international standard for laboratory competence. Accreditation bodies evaluate the lab’s quality management system, staff training, and the validity and traceability of its results. The credential is not one-time; labs face ongoing surveillance assessments to keep it.4A2LA. Cannabis Testing Laboratory Accreditation State rules often layer on top, with some states folding their own testing requirements into the accreditation assessment through a memorandum of understanding with the accrediting body.
Independence From Producers
Testing labs must be financially and operationally separate from the businesses whose products they test. A lab owner cannot hold a financial stake in a cultivation, manufacturing, or retail operation, and the lab cannot share employees with, lease space from, or carry any outside business relationship with licensees it tests for. The firewall exists because the incentives to produce favorable numbers are enormous when revenue depends on keeping clients happy.
Proficiency Testing and Inspections
Labs analyze blind reference samples with known concentrations through proficiency testing programs. Results outside the acceptable range flag problems with equipment calibration or analyst technique. Some states run their own proficiency programs; others require approved third-party providers. State inspectors conduct unannounced visits to review records, observe procedures, and verify the lab is following its own approved standard operating procedures. Consistent failures can result in fines, suspension, or permanent license revocation.
DEA Registration for Hemp
Federal law draws the line between legal hemp and federally prohibited cannabis at 0.3% Delta-9 THC on a dry weight basis. The 2018 Farm Bill defines hemp as any part of Cannabis sativa at or below that threshold.5Office of the Law Revision Counsel. 7 USC 1639o – Definitions Any lab testing hemp for THC concentration under the federal program must be registered with the Drug Enforcement Administration, because a hemp sample that tests above 0.3% is legally a controlled substance the moment the result comes back.6Agricultural Marketing Service. Hemp Analytical Testing Laboratories Labs serving only state-legal adult-use or medical markets operate under their state’s licensing framework, while the broader federal scheduling conflict stays unresolved.
When a Batch Fails
A failed test does not automatically send a batch to destruction. Most states give producers a structured path to try to save it. The lab notifies the producer of the failure, and the producer can usually request a reanalysis of the original sample within a short window, often seven days. If the reanalysis passes, a second independent lab may need to resample and confirm. If the failure stands, the producer can submit a remediation plan describing how the batch will be brought into compliance, backed by evidence the method will work. The state agency approves or rejects the plan, approved remediation must begin within a set timeframe, and the remediated batch must pass the full testing cycle again before release.
If remediation fails, the deadline is missed, or the plan is rejected, the batch must be destroyed. Destruction itself is regulated and tracked through the state system. One rule holds across jurisdictions: mixing a failed batch with a passing one, or remediating in an unapproved way, contaminates the whole combined quantity regardless of what a later test shows. Blending is not a way out of a failure.
What Testing Costs
A basic cannabinoid potency analysis typically runs $50 to $100 per sample. A full compliance panel covering potency plus pesticides, heavy metals, microbial contaminants, residual solvents, and moisture content runs $200 to $400. Batch size caps multiply the bill: a grower with 500 pounds of flower subject to a 50-pound batch limit pays for ten separate rounds, and any failure that requires retesting after remediation adds more. For smaller operators especially, testing is a significant and often underestimated line item.