Does Your Cannabis Match Its Label?

Key Takeaways

• Independent testing has found widespread label inaccuracy in both state-licensed cannabis and unregulated hemp CBD, though the two markets operate under very different rules and should not be lumped together.
• In a 2023 study of licensed Colorado dispensaries, about 70% of flower samples contained meaningfully less THC than the package claimed.
• Across four studies of hemp CBD products, accuracy rates ranged from 24% to 54%.
• Errors run in both directions, and products marketed as THC-free have been found to contain detectable THC.
• A batch-specific Certificate of Analysis is the only reliable way to learn what a product actually contains.

---

You buy the same product twice. Same brand, same label, same number on the front of the package. The first one does what you hoped. The second does almost nothing, or considerably more than you bargained for. If that has happened to you, the easy explanations are that you ate differently that day, or slept badly, or built up tolerance. Those things are all real and they all matter.

There is another explanation that gets discussed far less, and researchers have spent the better part of a decade documenting it. The two products may simply not have contained the same thing. When scientists buy cannabis and CBD products off the shelf and measure what is inside them, the number on the label turns out to be an unreliable guide, and not occasionally. Most of the time.

Two Different Markets, One Shared Problem

Before the numbers, a distinction that most coverage of this topic collapses and shouldn't.

State-licensed cannabis is what you buy at a dispensary. It faces mandatory testing requirements set by the state, moves through seed-to-sale tracking systems, and can be sold only to verified adults. Hemp-derived CBD is a different animal entirely. Federal law removed hemp from the controlled substances list in 2018, and the products that followed, the oils and gummies and creams stocked online and in ordinary pharmacies, face almost no mandated testing at all.

Most published research on label accuracy has examined the second category, because those products are easy to buy and ship. The result is an odd gap. The market with the least oversight is the one we know the most about.

The evidence below is separated accordingly, so you can tell which findings apply to what you are holding.

What the Testing Found in Licensed Dispensaries

The most recent look at this is also the most pointed. In 2023, researchers bought 23 flower samples from 10 licensed dispensaries across Colorado's Front Range and measured THC by laboratory analysis. Average measured potency came to about 15%, some 23% below the lowest figure printed on the packaging. Around 70% of samples fell more than 15% short of their labeled THC, and three contained close to half the maximum potency claimed [1].

The authors were careful about what they concluded, and their explanation matters as much as the number. They point to three structural drivers: no standardized testing protocol across laboratories, limited regulatory oversight of how results get produced, and a market in which price tracks THC percentage, giving everyone along the chain a reason to prefer the higher figure. What they describe is a system with badly aligned incentives. Deception at the counter is not the mechanism.

The study has two limits. Twenty-three samples from a single metropolitan region is small, and two of its four authors work for a cannabis testing and manufacturing company. Read it as a strong signal about one market at one moment.

Edibles have a longer track record and it is not better. In 2015, a team purchased 75 edible cannabis products from dispensaries in Seattle, San Francisco, and Los Angeles, spanning 47 brands. Only 17% were accurately labeled for THC. Sixty percent contained less than the package claimed, and 23% contained more [2].

What the Testing Found in Hemp CBD Products

The unregulated side has been studied more thoroughly, and the results hold steady across years and product types.

In 2017, researchers bought 84 CBD products online from 31 different companies. About 31% contained CBD within 10% of the labeled amount. Of the rest, 43% held more CBD than advertised and 26% held less [3].

Five years later, a University of Kentucky team tested 80 hemp-derived CBD products bought online and from local retailers. Just over half, 54%, came within 10% of the label claim, meaning nearly half did not [4]. That same year, a Johns Hopkins group examined topical products, the creams and lotions and patches stocked in pharmacies and grocery stores nationwide. Of the 89 products listing a CBD amount, 24% were accurate, and the typical in-store product deviated from its label by 21% [5].

Why the Errors Run Both Ways

Mislabeling would be easier to manage if it always meant getting shortchanged. It cuts both directions.

When a product contains less than the label says, the consequence is disappointment and wasted money. You take what should be an effective amount, feel nothing, and reasonably conclude that cannabis does not work for you. The dispensary flower and edibles studies both fell mostly into this category.

When a product contains more, the consequence is different and less pleasant. Someone titrating carefully upward, adding a small amount at a time to find a comfortable level, can be pushed well past it by a product that happens to be stronger than advertised. The hemp CBD studies mostly ran this way, with more of the compound present than the package promised.

A label number, then, is not a dose. It is an estimate of unknown reliability, and you cannot predict the direction of the error from the brand, the price, or the packaging.

The THC Nobody Mentioned

One finding reaches beyond how a product feels.

In the 2017 online CBD study, THC turned up in 18 of the 84 samples tested [3]. In the topical study, THC was detected in 35% of products. All of it sat under the 0.3% legal threshold for hemp, but among those THC-containing products, 11% were labeled THC-free outright and half made no mention of THC on the label at all [5].

For most people this is a curiosity. For anyone subject to workplace drug testing, on probation, or in a safety-sensitive job, it is not. A product bought specifically because it claimed to be free of THC may not be.

What the Label Leaves Out Entirely

A further gap comes from what labels omit.

Nearly every cannabis label reports THC and CBD. Almost none report terpenes, the aromatic compounds that give different cultivars their distinct smell. There is reason to think terpenes matter. In a controlled trial at Johns Hopkins, participants who inhaled THC together with d-limonene, the terpene behind citrus scent, reported feeling significantly less anxious and paranoid than on the same dose of THC alone. The effect appeared at one specific combination, 30mg of THC with 15mg of d-limonene, and only among a subset of twelve participants [6].

That finding carries limited weight on its own. A comprehensive 2024 review concluded that while terpenes plausibly shape how cannabis feels, the broader claim that they meaningfully enhance cannabinoid effects remains unproven and awaits proper clinical trials [7]. We examine that question in more depth in our Research article on the entourage effect.

The point holds regardless of how that science resolves. Two products with identical THC percentages can carry completely different terpene profiles, and the label will not tell you.

How to Check Before You Buy

None of this means shopping blind. The label is a starting point, and a better document exists.

A Certificate of Analysis, usually called a COA, is the laboratory report for a specific batch. Reputable producers make it available, often through a QR code on the package or a lookup tool on their website. Four things to check.

First, whether a COA exists at all. If a company cannot produce one, that tells you most of what you need to know.

Second, whether it matches what is in your hand. A COA is batch-specific, and a generic report posted once and never updated describes a product somebody else bought. Look for a batch or lot number matching your package.

Third, what it covers. Many COAs report only THC and CBD. A fuller panel adds minor cannabinoids, a terpene profile, and safety testing for pesticides, solvents, and heavy metals. More detail is a reasonable proxy for a producer taking the work seriously.

Fourth, how closely the numbers track the label. Researchers generally treat anything within 10% as accurate. A report sitting well outside that range on its own product should give you pause.

Beyond the document, the people behind the counter can help. Tell your dispensary staff what you are trying to accomplish and ask what testing they can show you. Someone who can pull up a current batch report is offering you something the package cannot.

The Bottom Line

The evidence here is unusually consistent for a young field. Five research groups, working across different product categories, different states, and different years, all found that most of what they tested did not match its own label. This is not the behavior of a few bad actors. It describes how the market has functioned through 2023, which is as recent as published testing goes.

What follows from that is modest. If a product surprises you, the product may be the reason. If you are titrating carefully, build in room for the possibility that the number on the front is approximate. If THC content matters for legal or professional reasons, verify it instead of trusting a marketing claim. And when a producer makes batch-level testing easy to find, that is a reason to choose them.

For a fuller, fully referenced treatment of why this matters for clinical research and what analytical standards the field needs, see our companion Hytiva Research article, Cannabis Product Standardization and Quality: Why Consistency Matters for Research.

Frequently Asked Questions

Does this mean cannabis products are unsafe? Not usually. Most of this research concerns accuracy, not safety. A product containing 20% more CBD than advertised is mislabeled but not dangerous. The safety question is narrower and applies mainly to undisclosed THC and to contaminants such as pesticides and heavy metals, which is why a safety panel on a Certificate of Analysis matters.

Are dispensary products better labeled than hemp CBD sold online? Not reliably, though the reasons differ. Licensed cannabis faces mandatory state testing that hemp CBD does not, yet the most recent dispensary study still found about 70% of flower samples falling well short of their labeled THC. Researchers attribute that to inconsistent testing protocols between laboratories and to a pricing system that rewards higher reported potency.

Why is this so hard to get right? Cannabis is an agricultural product, and its chemistry shifts with genetics, growing conditions, harvest timing, and extraction method. Testing laboratories also differ in method and standards, so two labs can return different numbers for the same sample. None of that excuses an inaccurate label, but it explains why consistency takes deliberate effort.

What does a 10% deviation actually look like? On a product labeled 10mg of THC, a 10% deviation means somewhere between 9mg and 11mg, which most people would not notice. The studies found many products far outside that range, and that is where the difference shows up.

Should I stop trusting labels entirely? No, but read a label as a claim awaiting verification. Producers who invest in accurate labeling and accessible batch testing do exist, and asking for a Certificate of Analysis is how you find them.