You bought two 10mg gummies, same brand, same bag. One did almost nothing. The next one flattened you. That gap isn't your imagination, and it isn't a tolerance fluke — it's measurable, it's been studied, and the "10mg" on the package was always more of an estimate than a promise.

This isn't a reason to panic. It's a reason to understand what you're actually holding.

Two different problems wear the same costume

When people talk about edibles being "off," they're usually describing one of two separate phenomena — and the difference matters.

Label drift is the gap between what the package advertises and what the product actually contains, on average. It's a problem with the number on the box.

Piece-to-piece drift is the variation between individual gummies in the same package. The bag might average out fine, but gummy #3 and gummy #7 are not the same dose.

One is a problem with the label. The other is a problem with the candy.

They have different causes, different fixes, and a smart consumer treats them as two issues, not one.

Label drift: the number on the box has a documented credibility problem

The foundational data here is old enough to be a teenager and still gets cited because nobody's beaten it. In 2015, Vandrey and colleagues published a study in JAMA (Vandrey et al., 2015, JAMA 313:2491–2493) that tested 75 edible cannabis products bought from dispensaries in three cities. Only 17% were accurately labeled — meaning within ±10% of the THC content claimed on the package. 60% were over-labeled (less THC than the label said) and 23% were under-labeled (more THC than the label said).

Ten years later, the picture hasn't snapped into focus. A 2025 study out of Mississippi (PMID 41025251) used gas chromatography–mass spectrometry on cannabis edibles and found measured THC ranging from 288mg less to 5,491mg more than the package claimed. Colorado data tells a similar story from a different angle: research on labeled THC potency across flower and concentrate products (Sci Rep, 2025) and a separate analysis of 74 commercial products (PMC11952622) both documented label-versus-observed mismatches, with over-labeling a recurring theme.

A decade of studies, three states, and the same finding: the label is a starting point, not a measurement.

Piece-to-piece drift: why THC won't behave in a gummy

Here's the part that surprises people. Even when a manufacturer nails the total dose for a batch, the individual gummies still don't match each other.

A 2025 LC-MS/MS analysis of gummy products (PMID 40454463) measured the variation between individual gummies pulled from the same bottle. The coefficient of variation — basically, how much the pieces scatter around their own average — ran 2.1% to 27.1% for CBD and 3.1% to 23.5% for Δ9-THC. The low end is tight. The high end means one gummy can carry roughly a quarter more or less than its sibling.

The reason is physical, not sloppy. THC is fat-soluble. A gummy is mostly water, sugar, and pectin — not fat. When you try to disperse an oil-loving compound evenly through a water-based slurry at production scale, it doesn't want to spread out uniformly. It clumps, it settles, it concentrates in some pours and thins out in others. Stirring helps. It doesn't make the problem disappear.

Make your own edibles and you sidestep the supply-chain version of this problem — but you inherit the dosing math yourself. Our cannabis brownies guide walks through the decarb-and-divide arithmetic that keeps a homemade batch honest.

THC in a gummy matrix is oil trying to behave in water. At scale, "evenly mixed" is an aspiration.

Age, heat, and light don't help

A gummy is not a stable object frozen at its production-day potency. THC degrades over time, and the process speeds up with heat and light. A bag that lived in a hot car or sat in a sunny window for a season is not the same product it was on the shelf. This works with label drift, not against it — an over-labeled gummy that's also been aging in a warm drawer drifts further from its number, not back toward it.

The dose is only half the equation

Even a perfectly labeled, perfectly homogenized gummy wouldn't land the same way every time, because the gummy is only one input.

Onset and intensity also depend on whether you've eaten — edibles taken on an empty stomach can come on differently than ones taken after a full meal — plus your individual metabolism and your current tolerance. Two people splitting the same gummy can have genuinely different experiences, and the same person can have different experiences on different days. The milligram number describes the candy, not the event.

The label measures the gummy. It can't measure you, your last meal, or your week.

Δ9, Δ8, and hemp-derived are not one category

"Edibles" is a wide tent, and the drift problem isn't evenly distributed under it.

Regulated dispensary products — the Δ9-THC gummies sold in licensed adult-use and medical markets — are subject to state testing requirements and batch tracking. Drift still happens, as the studies above show, but there's a compliance system, however imperfect, sitting behind the label.

Hemp-derived products, including a lot of what gets sold as Δ8-THC, occupy a much looser landscape. The Mississippi study's most extreme discrepancies — including that 5,491mg overage — came from Δ8 products, and three of them traced to the same manufacturer. Different regulatory pressure, different quality-control incentives, different drift profile. Treating a gas-station Δ8 gummy and a dispensary Δ9 gummy as the same risk category is a mistake. They're not.

What the COA is, and why it's your best tool

A Certificate of Analysis — COA — is the lab report for a specific batch of product. An independent or state-licensed lab tests a sample and documents what's actually in it: cannabinoid content, and usually a screen for contaminants like pesticides, heavy metals, and residual solvents.

Where to find it: regulated products often have a QR code or batch number on the packaging that links to the COA, or you can ask the dispensary directly. Reputable hemp-derived brands post COAs on their websites, searchable by batch number.

A COA isn't paperwork theater — it's the one document that tells you whether the number on the box survived contact with a lab.

Three things that actually cut down on surprises

1. Re-baseline with every new product. Different brand, different batch, even a different flavor from the same brand — treat it as an unknown. Many experienced consumers re-baseline at 2.5–5mg when they pick up something new, then adjust from there once they know how that specific product behaves. The point isn't a magic number; it's not assuming a new gummy matches your old one.
2. Wait out the full onset window before redosing. Edibles commonly take up to two hours to fully come on, because the THC has to move through your digestive system and liver before it reaches you. The classic mistake is taking a second gummy at the 45-minute mark because "nothing's happening" — and then both doses land together. Give it the full window.
3. Check the COA before the candy. Match the batch number, compare measured content to the label, confirm there's a real lab behind it. Two minutes of reading is the cheapest insurance against the label-drift problem.

If a gummy hits harder than you planned for

It happens, and it passes. The effects of an edible are temporary even when they're more intense than you wanted. Practical moves: find a calm, comfortable spot, stay hydrated, and let time do the work — the discomfort fades as the dose clears. You'll see CBD recommended as a counter to too much THC; the evidence on that is mixed and it's not a reliable off-switch, so don't bank on it as a rescue plan. Mostly, the move is to ride it out somewhere you feel safe.

The posture, not the moral

None of this is an argument against gummies. It's an argument for holding the number on the package a little more loosely than the package wants you to.

"10mg" is a manufacturing target filtered through a fat-soluble compound, a water-based candy, a supply chain, and a shelf life. Sometimes it's dead-on. Sometimes it drifts. The consumers who don't get caught out aren't the ones with the best math — they're the ones who treat every new bag as a question, read the COA, and give the edible the two hours it asked for. Know that going in, and the ride gets a lot more predictable.

Read the label loosely. Read the COA closely.