The More Ingredients a Sleep Supplement Lists, the Less You Know About Any of Them

Two bottles. One lists four ingredients. The other lists twenty, in a bigger font, with more words on the front. Most people reach for the second one, because twenty sounds like more care went into it.

We read 600 sleep supplement labels from the National Institutes of Health’s public label database to find out what actually happens as that ingredient list gets longer. The answer turned out to be a pattern, not a number.

The longer the list, the less it tells you

For every label we counted the active ingredients and checked whether each one had its own stated dose — or whether it disappeared into a blend where only the total is printed.

The climb is steady, which is what makes it a rule rather than noise. A twenty-ingredient sleep formula is not twenty pieces of information. It is closer to seven pieces of information and thirteen names.

Across the whole sample, 38.6% of all active ingredients — 1,557 out of 4,036 — are sold without a stated dose of their own.

Seventy-six products don’t state a single dose

The number that stopped us: 12.7% of the products we read — 76 of 600 — disclose no dose for any active ingredient at all. Not one.

The most extreme carried seventeen ingredients and zero doses. You can read that label front to back and learn only what is in the bottle, never how much of anything.

Among formulas with ten or more ingredients, only 27.8% disclose every dose.

The obvious objection, and why it doesn’t hold

The first reasonable pushback: traditional herbal formulas — Ayurvedic, Chinese — are multi-ingredient by design, and printing per-herb doses isn’t how that tradition works. Judging them by a Western clinical-dosing yardstick would be unfair.

Fair point, so we checked it. Among the 76 products disclosing nothing, recognisably traditional formulas account for 5%. Among all products using blends, 4%. The tradition explains a corner of this picture. The remaining ninety-odd percent are ordinary supplement brands.

The part that argues against the headline

Here is what we’re not going to bury: 58% of the products we read disclose every single dose. The majority of this category behaves properly.

That is the third time in a row our data has come back less damning than the story we could have told. We counted ashwagandha labels expecting systematic underdosing and found the doses were mostly fine. We counted melatonin expecting the category to be reckless and found something more specific than recklessness. Here, most brands print what they put in.

The problem is concentrated, not universal. That’s a less exciting sentence and a truer one.

None of this is against the rules

Proprietary blends are legal. A manufacturer may list the ingredients in a blend and print only the combined weight, and there is a real argument behind it: a formula that took work to develop can be copied off its own label in an afternoon.

We think that argument is weaker than it sounds — ingredient identities are printed either way, and what stays secret is only the ratio — but it isn’t nonsense, and pretending otherwise would be dishonest.

The cost, though, lands entirely on the buyer. A blend cannot be compared to a study, because a study tested a dose. It cannot be compared to a competitor, because their total is made of different proportions. It can only be compared to the marketing around it.

Where we have an interest, and where our data is blunt

This one deserves a harder disclosure than usual, because it flatters us more than usual. We make a sleep supplement with four ingredients and every dose printed on the label — along with the twelve ingredients we rejected and why. A finding that reads “short lists are more transparent” describes our own product. Weigh it accordingly.

So let us also say what this data does not show, including the part that undercuts us.

A disclosed dose is not automatically a meaningful one. A product can print fifteen numbers and every one of them can be a token amount. We measured whether a dose is stated, not whether it is worth stating — and by our measure such a product scores perfectly. Full transparency about a weak formula is still full transparency, and it is not the same thing as a good product.

A short list is not automatically better. Four ingredients is a choice, not a virtue. An eight-ingredient formula with eight honest doses would beat ours if the doses were better chosen.

And the counting involved judgement. We used the database’s own ingredient categories to decide what counts as active rather than an excipient. Someone drawing that line differently would get numbers a few points off ours. The pattern across the size bands is steep enough to survive that, but the individual percentages are not sacred.

What to do with this in a shop

One question does most of the work: can I see a number next to every ingredient I care about?

If a name sits inside a blend, treat that ingredient as absent from your decision. Not necessarily absent from the capsule — but you have no way to reason about it, so it cannot count toward why you’re buying.

And when a label lists twenty ingredients, ask what the twentieth is doing there. In a capsule the size of a capsule, the twentieth ingredient is usually present at an amount chosen so it can appear on the label.

How we counted, so you can check us

We searched the DSLD for “sleep”, filtered to products currently on the market, and read 600 labels on 29 August 2026. The database held 3,725 matching records, so this is a sample, not the whole shelf.

For each label we walked the full supplement facts panel, including ingredients nested inside blends. An active ingredient was counted as disclosed if it carried its own quantity, and hidden if it appeared only inside a blend total or carried no quantity at all. Excipients, flavourings and nutrition-panel lines — calories, sugars, fats — were excluded from the counts. Blends were only counted as proprietary when named as a blend of actives; standard phrases like “contains less than 2% of” were excluded.

The database is free and needs no account. The method is simple enough to repeat, and if you get different numbers we’d like to hear it.

†These statements have not been evaluated by the Food and Drug Administration. This product is not intended to diagnose, treat, cure, or prevent any disease. This article is educational and is not medical advice; if you have a medical condition or take medication, talk to your healthcare provider.

Update: we went back and measured exactly that. The follow-up study checks every stated dose in these 600 formulas against the trials — across four verified ingredients, 27.4% reach the studied dose.

The counter-example is instructive: L-theanine is the one ingredient this category doses honestly — zero hidden doses in standalone products, and only 10.8% hidden overall.

We later found the category’s worst offender: valerian hides its dose on 44% of labels — the highest rate of any ingredient we counted.

We also ran a different kind of count on this category: not doses but safety warnings — 8% of 5-HTP products carry no medication warning at all.