Proprietary Blends: The Legal Trick That Hides Doses

What the label actually tells you

Supplement Facts panels in the US are governed by 21 CFR 101.36. For most dietary ingredients the panel must state the quantity per serving. For a group of ingredients the manufacturer chooses to call a proprietary blend, the rule changes. Paragraph (c)(3) provides that “the quantitative amount by weight specified for the proprietary blend shall be the total weight of all other dietary ingredients contained in the proprietary blend,” and that those ingredients “shall be declared in descending order of predominance by weight.”¹

So a blend line gives you exactly three true facts:

• Every ingredient in the blend is named. Nothing may be omitted from the list.
• The combined weight of all of them, per serving.
• Their rank order by weight — first is heaviest, last is lightest.

That is the whole disclosure. It is more than nothing, and it is far less than most buyers assume they are reading.

What it doesn’t tell you — the legal gap

No individual dose. Descending order is an extremely weak constraint. Take a 500 mg “cognitive support blend” listing bacopa first, then eleven other botanicals. Bacopa could be 489 mg with the remaining eleven splitting 11 mg between them, or bacopa could be 45 mg with the rest at roughly 41 mg each. Both label the same. Both are legal. Only one of them contains a dose that has ever been tested.

No verification before sale. The NIH Office of Dietary Supplements states the position plainly: “Medicines must be approved by FDA before they can be sold or marketed. Supplements do not require this approval.”² The panel is a manufacturer’s assertion, not a tested result. A 2020 analysis in Neurology: Clinical Practice detected five unapproved drugs — omberacetam, aniracetam, phenibut, vinpocetine and picamilon — in US-sold cognitive-enhancement supplements, frequently at doses above pharmacological norms and often undeclared or mis-declared.⁴

Nothing about whether the dose is too high. Underdosing is the usual failure, but a blend conceals the opposite just as effectively. Vitamin B6 has a tolerable upper intake level of 100 mg a day, and sensory peripheral neuropathy has been reported at doses as low as 50 mg a day with prolonged use.⁵ Inside a blend total, you cannot tell whether you are taking 5 mg or 75 mg — and the ceiling is cumulative across every product you take, not per bottle.

The arithmetic that gives it away

You do not need the hidden numbers to reach a verdict. You need the studied doses and a calculator. These are the doses that appear in the trials, with the population attached — because a dose without a population is not a fact.

Ingredient	Dose in trials	Population	Duration
Bacopa monnieri	300–450 mg/day, ~50% bacosides	Healthy adults	≥12 weeks
Citicoline	250–500 mg/day	Healthy adults; more in impairment	Weeks to months
Alpha-GPC	1,200 mg/day	Dementia patients, mostly with donepezil	Months
Creatine monohydrate	3–5 g/day	Memory effect mainly in adults 66–76	Weeks to months
Ginkgo (EGb 761)	120–240 mg/day	Diagnosed dementia only	~6 months
L-theanine	~200 mg single dose	Healthy adults, acute only	1 hour
PQQ	20 mg/day	Small Japanese trials, mixed results	12 weeks

Now add three of them together at trial dose: bacopa 300 mg, citicoline 250 mg and alpha-GPC 1,200 mg is 1,750 mg before you reach the fourth ingredient. A typical “brain blend” serving is 500–900 mg in total. The gap is not a rounding error; it is the entire product.

Creatine makes the point most bluntly. Its cognitive dose is 3–5 grams, which is 3,000–5,000 mg. No capsule blend on the market contains it, and a blend that lists creatine among twenty ingredients in a 750 mg matrix is listing it for the label, not the effect. The doses each of these compounds actually used are set out per ingredient on the bacopa dosage page and the alpha-GPC dosage page.

How to check a label in two minutes

1. Find the blend line and its total. It will say something like “Proprietary Cognitive Matrix 620 mg.” That number is your budget for everything beneath it.
2. Count the ingredients underneath. Twelve names under a 620 mg budget averages 52 mg each.
3. Look up the trial dose of the first-listed ingredient. It is the heaviest, so it is the best case. If its studied dose is larger than the whole blend total, you are finished — no ingredient in that product reaches a tested amount.
4. Check the serving size, not the capsule. “Serving size: 3 capsules” means the printed figures apply to three, and a 60-capsule bottle is a 20-day supply. This is where per-bottle price comparisons quietly go wrong.
5. Check standardization, not just the plant. “Bacopa monnieri 300 mg” and “Bacopa monnieri extract 300 mg standardized to 50% bacosides” are different products. Whole-herb powder doses from older Ayurvedic practice are not equivalent to the standardized extracts used in the meta-analyzed trials.
6. Scan the blend list for red-list names. Their presence ends the evaluation regardless of the arithmetic.

What a good one looks like — criteria, not brands

Products change formulas; criteria do not. These are the ones we apply before a product is discussed anywhere on this site.

• A milligram figure next to every ingredient. Not a blend total. This is the single filter that removes most of the category.
• Standardization stated where it matters — bacosides for bacopa, flavone glycosides and terpene lactones for ginkgo, fruiting body versus mycelium for lion’s mane.
• At least one ingredient at a dose that appears in a published human trial, in a population resembling the buyer. A dementia-trial dose is not evidence for a healthy 30-year-old.
• A named legal entity and a physical address. Not a support email and a checkout page.
• Third-party verification, or an honest absence of it. An NSF or USP mark confirms that the contents match the label and that contaminants are within limits — it says nothing about whether the ingredient works.⁸
• No red-list ingredient at any dose, disclosed or otherwise.

Meeting all six makes a product evaluable. It does not make it effective. Not one compound in our reference file earns a strong evidence rating for improving cognition in healthy adults, and a transparent label on a weak ingredient is still a weak ingredient — see the tier list for where each one actually stands.

Whether the premium is worth it — cost per effective dose

The right way to compare two supplements is cost per effective dose: the monthly price divided by the number of trial-strength daily doses the bottle actually delivers. Work an example. If a product contains bacopa at 300 mg standardized to 50% bacosides, one serving is one effective dose, and a 30-serving bottle at any price gives you a number you can compare against a single-ingredient bacopa extract.

Now try the same calculation on a proprietary blend. The denominator is unknown, so the quotient is undefined. That is not a technicality — it is the verdict. You cannot establish that a blend is expensive, and you also cannot establish that it is cheap, because you cannot establish that it delivers anything. A cheap product that delivers no studied dose has an infinite cost per effective dose, which is a worse deal than a costly one that delivers a real one.

The practical consequence is that single ingredients almost always win this comparison, because they are the only form in which the sum is computable. Buying bacopa, creatine or L-theanine separately also lets you test one variable at a time, which a twelve-ingredient matrix makes impossible by construction.

Why the convention survives

The stated justification is trade-secret protection: a competitor could copy a formula from a fully disclosed panel. That argument has some force for a genuinely novel ratio of ingredients and very little for a blend of twelve botanicals that every competitor already sells. The ingredient names are public in either case; only the numbers are hidden, and the numbers are what a competitor would find hardest to reverse-engineer and a customer would find most useful.

The commercial function is simpler. A long ingredient list reads as thoroughness, an unfamiliar botanical name reads as sophistication, and a single large-looking total reads as potency. A blend lets a manufacturer sell all three impressions at once while spending the ingredient budget on whichever component is cheapest. It also insulates the product from exactly the kind of comparison this page describes. The same structure sits behind most video sales letters in this category: an origin story, an undisclosed blend, six-bottle bulk pricing and a refund window. A specific number of years of memory “reversed” is the version of that pitch regulators have already litigated over.⁹

What would change our mind

Two things. A rule change requiring per-ingredient quantitative disclosure would make this page obsolete overnight, and we would be glad of it. Short of that, a manufacturer publishing independent third-party assay results showing that its blend delivers a trial-strength dose of its headline ingredient would defeat the arithmetic argument for that specific product — the assay, not the marketing claim. We have not seen one in this category.

Frequently asked questions

Are proprietary blends illegal?

No. They are expressly permitted by 21 CFR 101.36(c)(3), which requires the total weight of the blend and a descending-order ingredient list and nothing more. The problem is not legality — it is that the disclosure the law requires is not enough to evaluate the product.

Can I work out the doses from the order of the ingredients?

Only in one direction. You know the first ingredient weighs at least as much as the second, and that no single ingredient can exceed the blend total. That is enough to disprove a claim — if the blend totals 500 mg, nothing in it reaches creatine’s 3 g cognitive dose — but never enough to confirm one.

Is a proprietary blend always a bad product?

It is always an unevaluable one, which for our purposes amounts to the same thing. A manufacturer confident that its doses match the published trials has a strong marketing reason to print them. Hiding a number that would help you is a decision, and it is reasonable to read it as one.

Does “clinically studied blend” mean the blend was studied?

Usually it means one ingredient in it has been studied somewhere, at some dose, in some population — not that this combination at these amounts was ever tested. Look for the study to be named, the dose to match the label, and the population to resemble you. If any of the three is missing, the phrase is doing marketing work rather than evidential work.

What about blends that are mostly vitamins?

Vitamins and minerals must be declared with their own quantities and percent daily values under the same regulation, so they usually sit above the blend line rather than inside it. That is worth noticing: on many “brain” labels, the only ingredients whose doses you can actually read are the cheap ones.

Related reading

• Nine ingredients that signal a bad brain supplement — the names that end an evaluation
• Do nootropics actually work? — the tier list, with the population behind every effect
• The case against brain supplements — the large trials that failed
• Bacopa monnieri dosage — where the 300 mg figure comes from
• Vitamin B6 toxicity — the dose a blend can hide
• How we rate evidence — what we accept as a source