31 %. 34 %. 28 %.

Today, such THC figures determine which cultivars are considered “premium”, which products can command higher prices, and which growers or producers are perceived as particularly capable. In the public eye, these numbers appear objective and scientifically validated – after all, they come from modern laboratory analyses.

Yet only a few in the scene ask a simple but crucial question: How exactly are these figures determined?

The short answer is: not as clear-cut as marketing would suggest.

In fact, THC figures result from a complex interplay of measurement techniques, statistical analysis, the plant's biological variability, and not least, economic interests. Modern laboratory methods such as HPLC analysis allow for precise measurements, but at the same time create methodological leeway that, under real market conditions, reliably leads to systematically elevated THC values.

In this article, I explore why published THC figures are often far less representative than many consumers, patients, and even professionals assume, and why these numbers have nonetheless become the central currency of the cannabis industry.

For consumers, THC percentages influence purchasing decisions and price acceptance.
For medical patients, they can affect dosage, expected effects, and treatment safety.
For breeders and producers, they determine market positioning, reputation, and financial success.

In short: Hardly any single metric shapes today’s legal cannabis market as strongly as the THC value.

All the more problematic is that this figure often does not represent an objective property of a cultivar or batch, but rather a technically determined approximation subject to systematic distortions. You often pay for a potency you never get. Anyone who wants to understand how this happens needs to look behind the scenes of laboratory analysis and at the mechanisms that make so-called lab shopping possible in the first place.

Lab shopping refers to the practice of sending identical samples to multiple laboratories and then publishing only the result with the highest measured THC value. No breeder or producer submits samples multiple times to show the lowest result.

From the producers’ perspective, this practice is understandable: higher numbers mean better placement in shops, higher prices, and greater visibility. It is also lucrative for the laboratories themselves, as more tests generate more revenue. In this way, both sides benefit.

This practice is still rarely discussed openly. To the outside world, the impression of objective, standardized science remains, while internally it is well known that different laboratories can produce noticeably different results from identical material (inter‑lab variation). – Inter-Lab Variation in the Cannabis Industry, Part I: Problem and Causes | Cannabis Science and Technology – Cannabis Industry News, Insights

To understand why this is possible, it is worth taking a closer look at the technical basis of these measurements.

Many people assume that a laboratory value represents a fixed, unchangeable figure, comparable to the weight of an object.

In analytical chemistry, however, this is an illusion. Every measurement is the result of a process that involves assumptions, calibrations, statistical models, and technical parameters. Added to this is the fact that cannabis is not a homogeneous industrial product, but a biological system with natural variability.

Even with the greatest care, deviations inevitably arise – not as errors in the traditional sense, but as an unavoidable feature of complex measurement processes.

The most commonly used standard today for determining cannabinoids is the so-called HPLC analysis (High-Performance Liquid Chromatography).

Fig. 1: HPLC system (Agilent 1200).
Photo: Kommando / Wikimedia Commons (CC BY-SA 3.0).
https://commons.wikimedia.org/wiki/File:Agilent1200HPLC.jpg

HPLC separates the chemical components of a complex mixture and then quantifies them. The device does not measure 'THC percentage' directly; instead, it first generates an electrical signal over time, which is recorded peak by peak in a chromatogram. The area under each peak correlates with the amount of the respective compound.

Fig. 2: Chromatographic performance criteria from a single peak.
Photo: Shulalevin / Wikimedia Commons (CC BY-SA 4.0).
https://commons.wikimedia.org/wiki/File:Chromatography_Performance_Criteria_from_One_Peak.jpg

The concentration is then calculated using a previously established calibration curve. This method is technically robust, but it is not an absolute direct value – rather, it is the result of a series of assumptions, integration algorithms, and calibration parameters.

Therefore, even the smallest differences in instrument settings, reference standards, integration methods, or column conditions can lead to varying results, even for the same sample.

Analysts often speak of systematic deviations in HPLC measurements in the single-digit percentage range, yet with raw samples, complex matrices, and varying laboratory practices, this can routinely vary by ±10 % or more. In inter-laboratory comparisons, internal studies have observed deviations of up to 30 % for identical samples – far beyond what end users would intuitively expect. – The Cannabinoid Content of Legal Cannabis in Washington State Varies Systematically Across Testing Facilities and Popular Consumer Products | Scientific Reports

Specifically, this means that a product labeled as 30 % THC could, with a 20 % relative deviation, actually contain only 24 %. With a 30 % deviation, the value drops to 21 % – a difference that artificially determines 'premium status' and the sales price in the market.

It is important to note: such variations are not necessarily 'wrong' in the sense of being erroneous, but are the result of different methods, sample preparation, and laboratory practices. However, the methodological levers that measurably influence the outcome are no secret within the industry.

Cannabis is not a homogeneous material. Cannabinoid and terpene distribution varies:

• between plants of the same cultivar
• between different buds of the same plant
• between the upper and lower parts of the plant
• between harvest times and growing conditions

Scientific studies also show significant differences in total THC content within the same batch, even when measured using the same analytical method. – Variability of total THC in greenhouse cultivated dried Cannabis | Scientific Reports

This means that a single sample can never be representative of an entire batch or cultivar. Yet it is precisely such individual values that are communicated as brand or cultivar metrics.

At this point, analytics and market dynamics intersect.

The market rewards high THC numbers.
Retailers sell products labeled '30 %+' faster.
Producers achieve higher margins.
And laboratory practices make it possible to systematically influence measurement results.

This has empirical consequences:

• In a large-scale study of 277 cannabis products from Colorado, it was found that the observed THC potency was often significantly lower than indicated on the label – for flower, frequently outside an accepted ±15 % tolerance, while concentrates generally matched more closely. – Accuracy of labeled THC potency across flower and concentrate cannabis products | Scientific Reports
• Earlier studies also found that in 70 % of tested samples, the observed THC potency was significantly below the declared value, in some cases by more than 30 %. – Uncomfortably high: Testing reveals inflated THC potency on retail Cannabis labels | PLOS One

These findings do not exist in a vacuum, but reflect both economic incentives and systemic limitations in the analytical and regulatory environment. Given this consistent body of evidence, a random accumulation of systematic overestimations is practically ruled out.

The following overview categorizes the measurement methods used in the cannabis industry and illustrates why THC values from different sources are not directly comparable.

The near-infrared (NIR) spectroscopy listed in the table is increasingly used in the cannabis industry as a rapid method, particularly for screening purposes and process control.

Unlike HPLC, NIR does not determine active compound concentrations directly. The method captures the spectral properties of the sample and uses statistical calibration models to derive estimated concentrations, which are trained in advance using extensive reference measurements (typically HPLC data).

The reliability of such systems therefore depends largely on:

• The quality and scope of the underlying training dataset
• The match of the current sample with the calibration conditions
• Instrument configuration and mathematical model

In practice, this means:
NIR is by far the least accurate of the quantitative methods listed here, as measurement errors, model uncertainties, and biological variability accumulate.

At the same time, the method has clear strengths: it is fast, non-destructive, and well suited for rough potency estimates, trend analysis, or categorizing batches.

Commercial systems, such as Purple Pro, are based on this approach and use pre-trained models to derive estimated THC values from the measured spectral data.

Technically, this is not a primary measurement of active compound concentration, but a model-based approximation derived from previous laboratory analyses.

Against this background, marketing claims of 'laboratory precision' or 'analytical accuracy' for such systems should be viewed critically. The increasing promotion of these devices as full substitutes for laboratory-based analysis further blurs the line between estimation and measurement.

When high THC values become the central selling point, priorities in breeding and selection also shift:

• Terpene profiles take a back seat
• Complex effect profiles are neglected
• Genetic diversity shrinks (bottleneck)
• Medically relevant traits lose importance

In the long term, a market emerges that is not optimized for holistic quality, but for a single number.

I have been working with cannabis genetics, selection, and breeding for many years. During this time, I have seen countless analyses – and just as many THC figures that would never have been reproducible under realistic conditions.

For this very reason, I deliberately do not publish standardized cannabinoid or terpene content figures on my website.

Not because I consider this data inherently worthless, but because I believe it is dishonest to present single measurements as objective characteristics of a cultivar.

When analytical data is published at all, it is presented for what it is: a snapshot under clearly defined conditions, referring to specific plants, not as a universal quality guarantee. Economically, this decision is certainly not optimal – high numbers sell better. In the long term, however, I consider transparency more valuable than short-term marketing effects. Trust is built not through peak values, but through traceability.

THC figures are not useless.

They allow for:

• A rough classification of potency within a system
• Comparison within the same laboratory
• Guidance for dosage estimations in a medical context

What they do not show:

• An objective cultivar evaluation
• A reliable quality assessment across markets
• Absolute, constant active compound values

Anyone reading THC figures should therefore understand them for what they are: technically derived approximations – not absolute truths.

Modern analytics are powerful. HPLC is a precise, well-established method. Yet even the best technology cannot eliminate biological variability, methodological differences, and economic incentives.

As long as high THC values determine market opportunities, lab shopping will occur.
As long as single samples are sold as cultivar values, distortions remain unavoidable.
And as long as numbers serve as a substitute for quality, true transparency will remain the exception.

Published THC values are inflated across large parts of the market and are, in practice, misleading for consumers, as a significant portion of the values communicated in the market is systematically biased upwards according to current studies.

A critical approach to these figures is therefore not distrust of science, but an expression of scientific understanding and consumer awareness.

As long as THC values are communicated without methodological transparency, uncertainty information, or standardization, they are not an objective quality metric, but a systemically biased selling point – and it is up to the industry itself to finally scrutinize this practice critically.

• THC values are measurements, not natural constants
• HPLC is precise, but not free from systematic leeway or errors
• Single samples are not representative of entire cultivars or batches
• Lab shopping is real and economically motivated
• High numbers do not automatically mean high quality
• Transparency is more valuable than peak values

Giordano, G., Brook, C. P., Ortiz Torres, M., et al. (2025). Accuracy of labeled THC potency across flower and concentrate cannabis products. Scientific Reports, 15, 20822. https://doi.org/10.1038/s41598-025-03854-3

Schwabe, A. L., Johnson, V., Harrelson, J., & McGlaughlin, M. E. (2023). Uncomfortably high: Testing reveals inflated THC potency on retail cannabis labels. PLOS ONE 18(4): e0282396. https://doi.org/10.1371/journal.pone.0282396

Jikomes, N. & Zoorob, M. (2018). The Cannabinoid content of legal Cannabis in Washington state varies systematically across testing facilities and popular consumer products. Scientific Reports, 8, 4519. https://www.nature.com/articles/s41598-018-22755-2

Cleary, B. (2025). Variability of total THC in greenhouse cultivated dried cannabis flower. Scientific Reports. https://www.nature.com/articles/s41598-025-06962-2

Smith, B. C. (2019). Inter‑Lab Variation in the Cannabis Industry, Part I: Problem and Causes. Cannabis Science and Technology. https://www.cannabissciencetech.com/view/inter-lab-variation-cannabis-industry-part-i-problem-and-causes