How to Read a Protein Powder Label

Protein powder labels are designed to be difficult to compare. Once you know the handful of tricks in use, most of the confusion disappears and you can work out in about thirty seconds whether a tub is worth the money.

First, notice which panel you are reading

A protein powder can be sold as a conventional food, which carries a Nutrition Facts panel, or as a dietary supplement, which carries a Supplement Facts panel. The disclosure rules differ. Supplement Facts permits proprietary blends and “daily value not established” symbols; a Nutrition Facts panel does not work the same way. Which panel a brand chose is itself a piece of information, and it is on the back of every tub.

The protein number is a nitrogen estimate, not a protein measurement

This is the single most important fact on the label and almost nobody knows it. The standard method measures total nitrogen and multiplies by 6.25. It assumes that non-protein compounds contribute no nitrogen, and that essentially all the nitrogen present is protein. Both assumptions are exploitable, and that is the whole basis of the trick described further down.

There is a quality-corrected number, and its presence or absence is a useful tell. The percentage daily value for protein must be corrected for digestibility and amino acid quality — the PDCAAS adjustment. But it is only required when a manufacturer makes an explicit protein claim, or on foods for children under four. So: if a label shows a %DV for protein, someone has done the quality correction. If it shows grams only, nobody has. Whey sits at or near the top of the PDCAAS scale, so a genuine whey product loses nothing by disclosing it — which makes the omission considerably more interesting on a plant blend.

If you are wondering whether anyone has challenged this in court, they have. In Nacarino v. Kashi in 2023, the Ninth Circuit unanimously held that FDA rules expressly permit stating protein quantity by the nitrogen method on the front of a pack, provided the corrected %DV appears in the panel when it is triggered. The court described this as the agency striking a balance that keeps costs low for manufacturers. Which is to say: the loophole is not an oversight.

Start with the serving size, not the front of the tub

“25 grams of protein!” means nothing until you know what it is 25 grams of protein per. Some products quote per two scoops. Compare per 100 grams of powder instead — that is the only number that lets you compare across brands honestly.

An isolate should be around 80 to 90 grams of protein per 100 grams of powder. A concentrate is roughly 70 to 80. If it is 60, you are buying a great deal of something other than protein.

Three more serving-size forensics worth thirty seconds each:

Scoop weight versus serving weight. The panel states a serving in grams. The scoop may not deliver it, and an instruction to use a “heaping” scoop is a weighing problem rather than a nutrition fact.
“Servings per container: about.” Multiply servings by grams per serving and compare with the net weight on the tub. Discrepancies are common, and the arithmetic tells you how many servings you are really buying.
Per 100 grams is the only honest comparison basis. Convert everything to it before comparing two products.

The word “blend”

“Whey protein blend” almost always means a mix of concentrate and isolate, weighted towards the cheaper concentrate. The front of the tub will emphasise isolate; the ingredients list, which is ordered by weight, will tell you the truth. Read the ingredients, not the marketing panel.

Proprietary blends

The rule is specific. A proprietary blend must be identified as such, its total weight must appear on the same line, and its constituents must be listed in descending order of predominance underneath. What is not required — ever — is the amount of any individual ingredient inside it. That is not a gap in the regulation. That is the purpose of the structure.

Which means a blend can be 98% the cheapest ingredient and 2% the one pictured on the front of the tub, and the label is entirely compliant. Descending order tells you that ingredient one outweighs ingredient two, and nothing whatsoever about the gap between them.

A manufacturer proud of its formula publishes the amounts. Treat proprietary blends as a reason to move on.

Amino spiking

Standard protein testing measures nitrogen and infers protein from it. Cheap free-form amino acids — glycine, taurine, alanine, lysine, creatine — are nitrogen-rich and register as protein on that test without contributing usefully to muscle protein synthesis. They work particularly well for this because they are small molecules, so nitrogen makes up a large fraction of their molecular weight, and they cost a fraction of what whey costs.

The most-cited case is the class action filed in 2014 against Body Fortress, where the complaint alleged that 8.5 grams of a claimed 30 grams of protein — more than a quarter of it — came from added free-form amino acids. The judge denied motions to dismiss. Related actions named other brands and store-brand products, and the wave ran through about 2016.

One correction to a claim you will see repeated: amino spiking is largely not illegal. FDA labelling regulations do not specify what should and should not count toward protein content. Two trade bodies, CRN and AHPA, issued voluntary guidance recommending that only peptide-bonded amino acid chains be counted. Voluntary is the operative word.

And an honest note on prevalence: we could find no recent systematic survey of the market. The litigation wave has not obviously recurred. But the incentive is unchanged, the regulation is unchanged, and the detection method is unchanged, so the right position is neither “it is solved” nor “it is rampant” — it is that the tell is still visible on the label.

The tell. Look for free-form amino acids high in the ingredients list of a product marketed purely as protein — especially glycine, taurine, alanine, creatine, or a separately named “amino acid blend”. Peptide-bonded whey already contains all of these. Adding them separately has no nutritional rationale but does raise the nitrogen reading. Then do the arithmetic: divide protein grams by scoop grams. A 30-gram scoop claiming 30 grams of protein is not impressive, it is impossible.

Sugars, sweeteners and what the panel actually reports

“Added sugars” has a legal definition: sugars added during processing, sugars sold as such, sugars from syrups and honey, and sugars from concentrated fruit or vegetable juice. It excludes sugars occurring naturally in milk, fruit and vegetables. That matters for whey, because the lactose in a concentrate is a naturally occurring milk sugar — it appears under total sugars but not under added sugars. A clean product can legitimately show sugar grams. Check which line they are on. Our piece on how much sugar is too much covers the daily numbers.

On the sweeteners themselves, two things get overstated in opposite directions.

A 2023 Nature Medicine paper linked higher blood erythritol with cardiovascular events in cardiac-risk cohorts, and showed increased platelet reactivity in laboratory work. It is worth knowing about and it is not a reason to panic: it is observational epidemiology plus mechanism, not a controlled outcome trial, and the ingestion arm involved eight people. “Erythritol causes heart attacks” is not what that paper says.

The WHO issued a guideline in 2023 recommending against non-sugar sweeteners for weight control or reducing disease risk. Read the fine print: the recommendation is conditional, and WHO itself notes the observed associations may be confounded by who chooses diet products in the first place. It is not “the WHO says sweeteners are dangerous”. It is closer to “they do not deliver long-term fat loss”. Note also that stevia and monk fruit are inside the scope of that guideline — “natural sweetener” is a marketing category, not a toxicological one. More on this in what the evidence says about artificial sweeteners.

Gums, lecithin and emulsifiers, reported honestly

This is where supplement writing is worst, in both directions. Here is the actual state of the evidence.

A 2025 placebo-controlled randomised trial in Clinical Gastroenterology and Hepatology put 60 healthy participants through an emulsifier-free run-in and then four weeks of one of five arms: carboxymethylcellulose, polysorbate-80, carrageenan, soy lecithin, or a rice starch control. Microbial diversity held steady. Short-chain fatty acids fell across the emulsifier groups relative to placebo. The carrageenan arm showed increased intestinal permeability at a single marginal p-value in a group of about twelve people. There were no significant differences in faecal calprotectin, CRP, cholesterol or serum inflammatory markers, and the authors concluded only that the results suggest a potential benefit from limiting emulsifiers.

Two corrections follow. First, the alarming animal literature on carrageenan is overwhelmingly about poligeenan, a deliberately acid-degraded low-molecular-weight compound that is not food-grade carrageenan and is not a permitted food additive. Blogs conflate the two constantly. Second, there is no evidence base for the common storefront claim that sunflower lecithin is healthier than soy lecithin. Sunflower lecithin exists to avoid a soy allergen declaration and to satisfy non-GMO positioning, which are reasonable commercial reasons and not nutritional ones.

The bottom line: these ingredients are there for texture and solubility, the human evidence shows measurable microbiome and short-chain fatty acid shifts and no demonstrated clinical harm at realistic intakes. Preferring to avoid them is a defensible preference. A brand calling them toxic is overselling a four-week trial in sixty people.

Which words on the front have legal definitions

“Organic” — real and statutory, backed by the USDA National Organic Program. This is the contrast case that makes the rest legible.
“Grass-fed” — no federal standard. USDA withdrew its grass-fed marketing claim standard in January 2016, and claims are now approved case by case on label submission. The one exception with teeth is Certified Grassfed by AGA, which requires grass and forage only from weaning to harvest, no antibiotics or hormones, pasture-based management, US birth and traceability, and independent inspection at least every fifteen months. See what the grass-fed difference actually amounts to.
“Natural” — not regulated. FDA operates an informal policy that nothing artificial or synthetic has been added that would not normally be expected in that food, and explicitly says it does not address farming practices, processing methods, genetic engineering or nutritional benefit. A comment period closed in 2016 and no rule has issued.
“Non-GMO” — a private third-party standard, not a government one. Worth knowing: whey from grain-fed cows is routinely verified, because the protein contains no detectable transgenic material regardless of what the cow ate.

Third-party testing: which logos mean something

NSF Certified for Sport screens for over 280 substances banned in sport and is the only independent programme recognised by USADA, MLB, NHL and the CFL. It sits on top of NSF/ANSI 173.
NSF/ANSI 173 is the quality mark underneath: label verification that what is on the label is in the tub, toxicology review of added ingredients, and contaminant testing, maintained by annual audits and periodic retesting.
Informed Sport tests every single batch before release. Informed Choice is the lower tier, testing periodically rather than every batch. Informed Protein separately verifies protein content, which is directly relevant to the spiking problem above.
USP Verified certifies declared ingredients at declared potency, absence of harmful levels of specified contaminants, disintegration, and manufacture under current good manufacturing practice — but does not screen for substances banned in sport.
Logos that mean nothing on their own: “lab tested” or “third-party tested” without a named certifier; “GMP certified”, since compliance is legally required rather than an achievement; “made in an FDA-registered facility”, which is a mailing-list function, not an approval; and any seal the brand invented itself.

The test that settles it: can you look this specific product up in the certifier’s public database? If not, the logo is decoration.

Heavy metals, and a conflict we should declare

Consumer Reports tested 23 powders and ready-to-drink shakes in late 2025, using its own concern threshold of 0.5 micrograms of lead a day, derived from California’s Proposition 65. Plant-based products averaged roughly nine times the lead of dairy-based ones. The worst performers came in several micrograms per serving; several whey isolates were at a fraction of a microgram or not detectable.

Now the context Consumer Reports itself supplies, which rarely survives into the headlines: the average US adult already takes in about 5.3 micrograms of lead a day from ordinary food, and the FDA’s interim reference level is 8.8 micrograms a day. Prop 65’s 0.5 is a warning-label trigger, not a harm threshold. For scale, half a cup of cooked spinach carries around 0.9 micrograms and would require a warning if whole foods were covered. They are not.

A separate January 2025 survey of 160 products reported that 47% exceeded Prop 65 thresholds, with organic products averaging around three times more lead than non-organic. That one has been fairly criticised for not being peer-reviewed, for incomplete methodology, and for a conflict in that the testing organisation also sells certifications.

Here is our conflict, stated plainly: this finding favours us, because we sell whey. The contamination story is overwhelmingly a plant-protein story — rice, pea and cocoa concentrate metals from soil, and milk does not. What a brand should not do is translate “lower than plant protein” into “safe” or “zero”, and what it can do that genuinely helps you is publish per-lot certificates of analysis with the assay method and detection limits. Disclosure is the differentiator. Ask for it, from us or from anyone.

What is actually worth checking

Protein per 100g of powder. The single most useful comparison number.
Isolate or concentrate, stated plainly. If it will not say, assume the cheaper one.
A %DV for protein. Its presence means someone applied the quality correction.
Length of ingredients list. Protein, sweetener, flavour, maybe an emulsifier. That is a good label. Twenty items is a formulation optimised for cost and mouthfeel.
Third-party testing you can verify in a database.

What is mostly noise

Hydrolysed protein, unless you have a specific clinical reason. Added BCAAs, which are already in the whey. “Digestive enzyme blends” in a protein isolate that contains almost no lactose. Elaborate flavour names. Anything described as “advanced”, “elite” or “pro”.

On price

The formula is simple: price divided by servings per container times grams of protein per serving. Then cross-check servings against net weight, because that is where the number quietly moves.

A second metric worth learning is protein density — grams of protein divided by grams per scoop. It catches both amino spiking and filler-heavy blends in one calculation.

And for ready-to-drink products, run the same maths. The gap versus powder is usually three- to six-fold. That is a legitimate convenience premium rather than a scam, but you should see the number before you decide it is worth paying.

A cheaper tub at 60% protein with a spiked amino profile is not cheaper. It is the same trick as a low sticker price with a large delivery charge.