Cannabis Potency Testing Inconsistencies: Why State Lab Results Vary 20%+ — and What Your Brand Can Do About It

A cannabis concentrate submitted to three different state-licensed labs came back with THC values of 68.2%, 74.9%, and 81.3% — all from the same homogenized sample. That’s a 13-percentage-point spread. In a market where shelf tier, retail price, and legal compliance all hinge on potency numbers, a 13-point swing isn’t an analytical curiosity. It’s a quality control failure. And based on what inter-laboratory studies keep showing, it’s far more common than most brands want to acknowledge.

Inter-laboratory variability in cannabis potency testing has been documented in peer-reviewed literature, state compliance audits, and proficiency testing data for years now. The problem persists not because labs are negligent, but because the regulatory frameworks governing cannabis testing — which exist at the state level, since federal scheduling still leaves testing oversight fragmented — leave significant methodological latitude. The result is a market where two fully licensed labs can return dramatically different certificates of analysis from the same product, and where “lab shopping” has evolved from an open secret into a standard procurement strategy.

For brands operating in California, this has moved well past a marketing nuisance. As the Department of Cannabis Control (DCC) expands its compliance activity and Metrc’s track-and-trace system generates an auditable paper trail on every tested unit, potency numbers that don’t hold up under scrutiny tend to become enforcement problems later — not just reputational ones.

The Scale of the Problem — and What the Research Actually Shows

The most thorough published analysis of cannabis potency inter-laboratory variability comes from a study circulated in Forensic Chemistry that sent hemp flower and concentrate samples across dozens of labs. The mean relative standard deviation (RSD) for total THC measurements across labs was approximately 22%. To put that in practical terms: if a product’s true THC content is 20%, a 22% RSD means the distribution of results across labs could plausibly run from roughly 15.6% to 24.4% — same plant, same sample day, completely different numbers.

California’s DCC has flagged inter-laboratory comparability as an ongoing concern in its regulatory guidance. As of mid-2026, the state has issued active licenses to more than 130 cannabis testing laboratories — more than any other US state. That density creates direct competitive pressure on labs to differentiate on turnaround time and price rather than on analytical rigor. When higher potency numbers demonstrably generate repeat business from brand clients, the incentive to apply stricter method controls — which sometimes produce more conservative or lower-end results — becomes a quiet competitive disadvantage.

Researchers tracking reported average THC content in California flower products since recreational legalization have noted a consistent upward drift each year, at a rate that significantly outpaces what selective breeding timelines alone would explain. That gap between biological reality and reported numbers is a data signature of systematic variance rather than genuine potency improvement across the market.

For brands, the practical risk isn’t abstract. DCC compliance data from 2022–2023 showed that a meaningful portion of retested cannabis products returned potency values more than 10 percentage points different from their original COAs. Products caught in that gap face market withdrawal requirements, label correction, and — depending on the deviation — potential license action against the associated retailer or brand.

Why State Licensing Doesn’t Standardize Results

Being DCC-licensed means a lab meets California’s baseline requirements under 4 CCR §§ 15000–15012: qualified analysts, calibrated and maintained instruments, documented chain of custody, and a validated scope of testing. Those requirements matter, and they weed out genuinely unqualified operators. But they don’t specify a single, mandatory analytical method for cannabis potency — and that gap is precisely where most of the variability lives.

The two dominant analytical approaches are HPLC (high-performance liquid chromatography) and GC (gas chromatography). Both can return accurate results under well-controlled conditions. But they handle the THC/THCA relationship differently in ways that directly affect the reported number.

In HPLC analysis, THCA and delta-9-THC are measured separately as distinct peaks, then mathematically combined using a decarboxylation conversion factor: Total THC = delta-9-THC + (0.877 × THCA). In GC analysis, heat at the injection port converts THCA to delta-9-THC before chromatographic separation — the method effectively measures post-decarboxylation THC without a separate conversion step. How a lab applies that 0.877 conversion factor, and whether its label claim calculation aligns with the chosen measurement approach, introduces a first layer of variability even before sample preparation is considered.

Sample preparation compounds this substantially. Particle size after grinding, choice of extraction solvent, extraction duration, and dilution factor all affect analyte recovery. Small, undocumented differences in sample preparation protocol between labs produce real differences in measured concentration — even when the analytical instrument itself is functioning accurately and within calibration.

AOAC International’s Official Method 2018.11 provides a reference procedure for cannabis potency analysis by HPLC-UV, and ASTM International’s D8196 standard covers flower potency as well. Both are voluntary frameworks. No current California regulation requires cannabis testing labs to adopt either, and many labs continue to operate on in-house validated methods that meet DCC’s general validation criteria but diverge meaningfully from the reference procedures.

What ISO 17025 Accreditation Actually Requires — and Why It Changes the Accountability Structure

ISO/IEC 17025:2017 accreditation from a recognized body — A2LA, PJLA, and Perry Johnson Laboratories Accreditation are the most common for cannabis labs operating in the US — demands something that state licensing doesn’t: continuous, externally verified performance.

The key mechanism is mandatory proficiency testing (PT). Under ISO 17025, accredited labs must participate in PT programs where they receive blind samples from third-party providers and return results within predefined acceptable ranges. Results are tracked. A lab that repeatedly returns z-scores outside ±2 on cannabis potency PT rounds — meaning its results fall more than two standard deviations from the consensus value — must initiate documented corrective action. Continued failure risks loss of accreditation. The accountability loop is direct and consequential in a way that the DCC licensing renewal cycle simply isn’t.

The second meaningful difference is measurement uncertainty reporting. Under ISO 17025, labs are required to estimate and disclose the uncertainty associated with each measurement. A certificate of analysis that reports “THC: 24.3% ± 1.8% (k=2, 95% confidence interval)” is communicating something categorically different from one that reports “THC: 24.3%.” The first tells the recipient exactly how much analytical confidence to place in that number. The second doesn’t — and most cannabis COAs issued in California don’t include it, because state regulations don’t require it.

PT data from Cannabis Proficiency Testing, Inc. and comparative inter-laboratory studies consistently show that ISO 17025 accredited labs achieve tighter agreement on cannabis potency measurements than non-accredited labs. Accreditation isn’t a guarantee of error-free results — no analytical method is — but the accountability infrastructure is structurally different. As of mid-2026, California doesn’t mandate ISO 17025 accreditation for cannabis testing labs. Oregon has moved in that direction; California’s DCC has explored it. Until a mandate exists, accreditation remains a voluntary quality signal that brands need to actively seek out.

Three Questions to Put to Your Cannabis Testing Lab Before the Next Submission

The practical question for any brand isn’t whether your lab is licensed — they all are, or they shouldn’t be operating. The question is what quality framework sits behind that license. Three specific questions cut to the core of it.

First: Which reference method does your potency analysis follow? If the answer is “our in-house method,” ask to see the method validation summary, specifically the precision and accuracy data against certified reference materials. An in-house method isn’t inherently problematic — but a lab that can’t produce validation documentation on request is one you should think carefully about.

Second: Do you participate in a proficiency testing program for cannabis potency? Ask for the most recent PT report. Each analyte should show a z-score within ±2. A z-score outside that range, without accompanying documented corrective action, is a meaningful flag. Reputable labs are not defensive about sharing this data — it’s a selling point, not a liability.

Third: Do your COAs report measurement uncertainty? If not, ask why. Under ISO 17025, it’s a requirement. Under DCC regulations, it’s not — which is why its absence from most California cannabis COAs isn’t technically a violation, just an information gap that lands on the brand when results get challenged.

None of these questions are adversarial. A well-run testing program will answer all three without hesitation and will likely welcome the conversation. A lab that can’t — or that responds defensively — is providing useful information about its quality culture regardless.

The cannabis industry spent its first decade of legalization building consumer trust around THC percentages that, in many cases, weren’t analytically defensible. The brands that establish their quality documentation now, while California’s enforcement posture is still maturing, will be in a categorically different position when scrutiny intensifies. That starts with understanding what your testing data actually says — and being honest about what it doesn’t.

---

Written by Nour Abochama, Vice President of Operations, Qalitex Laboratories. Learn more about our team

Talk to our team about your testing needs. Contact us

Related from our network

• Analytical method validation under FDA and ICH guidelines — Aurora TIC covers regulatory-grade method validation requirements for pharmaceutical and device manufacturers navigating FDA compliance.
• Third-party COA verification and supplier qualification for raw ingredients — Ayah Labs helps global buyers independently verify ingredient quality and certificate of analysis accuracy through accredited chemical analysis.