How supplement ads inflate results with percentages

Supplement ads inflate results mainly in three ways: they quote a relative percentage ("50% lower risk") without the absolute numbers behind it, they report improvement from the starting point without a placebo or control group, and they pick the single best number from a study. A big percentage tells you how large a change looks. It does not tell you how many people it affected, or whether the product caused it.

This guide shows the arithmetic, what the evidence says about why these numbers persuade, and a short checklist for any percentage on a sales page. It is general information, not medical advice. A doctor or pharmacist can help you judge what a number means for you.

What is the difference between relative and absolute change?

The Cochrane Handbook explains it with a simple case. A treatment that reduces the risk of death from 4% to 3% can be described as reducing the relative risk by 25%, or as reducing the risk by 1 percentage point. Both statements describe the same benefit. One sounds large and the other sounds small.

• Relative risk reduction: the change as a share of the starting risk. The Handbook notes that a risk ratio of 0.75 translates to a relative risk reduction of 25%.
• Absolute risk reduction: the plain difference between the two groups, here 1 percentage point, or "10 out of 1,000".
• Number needed to treat (NNT): how many people must take something for one to benefit. The Handbook discusses NNT alongside the other two.

Here is a made-up illustration, not drawn from any real product. Suppose 2 in 1,000 people in a placebo group have a problem over a year, and 1 in 1,000 in the supplement group do. An ad can truthfully say "50% lower risk". The absolute difference is 0.1 percentage points, and about 1,000 people would need to take the product for a year for one to benefit. The percentage is accurate and the impression it creates is not.

The Handbook adds a caution that helps with interpretation: even if relative effects are similar across groups, absolute effects differ according to baseline risk. A "30% reduction" matters much more for a group where the problem is common than for a group where it is rare.

Why do ads prefer the relative number?

Because it works. In a 1992 survey published in the Annals of Internal Medicine, Naylor and colleagues gave 100 internal medicine and family medicine clinicians the same trial results, either as absolute event numbers or as relative risk reductions. Ratings of effectiveness were higher for the group that saw relative reductions. The authors concluded that clinicians' views of drug therapies are affected by the common use of relative risk reductions in both trial reports and advertisements.

If trained clinicians are swayed by the format, shoppers reading a landing page are too. The Cochrane Handbook says the same about presentation in general: users of reviews are liable to be influenced by the choice of statistical presentation, and whole-number "natural frequencies" such as "20 out of 1,000 dropping to 10 out of 1,000" are best understood by consumers.

So when an ad gives you a percentage, convert it yourself. Ask "percent of what?" and "how many people out of 100 or 1,000?" If the page does not say, the number is not ready to be used.

What is the "no control group" trick?

A second favorite is the before-and-after percentage: "participants improved by 40% in 12 weeks". It sounds like proof. The FTC's health products guidance explains why it is not. It says human clinical studies should have both a treatment group and a control group, and that improvements over time in the treatment group alone could come from a placebo effect, spontaneous changes in health, or the "practice effect" on a test.

The FTC guidance gives a concrete example. A study of an osteoarthritis shake reported a statistically significant improvement in symptoms from baseline to day 90. The FTC says that result does not substantiate a claim, because it does not compare the treatment group's improvement with the control group's. In that example both groups improved, and the difference between them was not significant.

The same guidance flags a related point that gets lost in ads: a statistically significant result must also translate to a clinically meaningful benefit, because some significant results are too small to have real consequences for health. A percentage can be real and still be tiny.

Which other percentage tricks should you watch for?

• Association turned into a promise. "People who eat more of nutrient X are 38% less likely to be overweight" describes a pattern in survey data, not a controlled trial of a supplement.
• The best arm of a study. A number reported "at the highest dose" or "in the best responders" is not the average result, and it may not match the dose in the bottle.
• Surrogate numbers. A change in a lab value or a blood pressure reading is not the same as fewer heart attacks, and a percentage change in a lab marker can sound dramatic while the reading stays in the same range.
• Subgroups. The FTC guidance warns that post hoc subgroup analyses may suggest areas for future research but generally do not provide reliable evidence.
• Many outcomes, one headline. In the FTC's example, one of eight outcome measures reached significance and the other seven did not. The more outcomes tested, the greater the chance of a false positive.
• Price percentages. "Save 60%" depends entirely on the "regular" price it is measured against.

How do offers we audited use numbers?

These examples repeat what the reviews found. They are not new accusations.

• The BioVanish review notes that the page cites a line that people who consume more B vitamins are 38% less likely to be overweight. The review points out that this describes an association in dietary intake data, not a controlled trial of supplemental megadoses. The same review compares the 5 g of MCT powder per scoop with the 18 to 24 g of MCT oil a day used in the cited weight-loss trial.
• The Cardio Shield review covers a page that cites a 7.2 point drop in systolic blood pressure from hibiscus and an 11 point drop from olive leaf at the highest dose. The review says the hibiscus figure is roughly consistent with trials, but the doses cannot be checked because six plant ingredients share one 570 mg blend with no individual amounts.
• The Blood Pressure Program review says published mind-body studies generally show reductions of single digits to the teens in mmHg on average, which the review calls a real but modest effect and not a reliable substitute for medical treatment. It also points out that the page's testimonials about dropping medication sit in tension with its own FAQ, which says not to stop medication without a doctor.
• The FLOW Nitric Oxide Booster review notes a "Save 25%" subscription price of $22.47 a month that disagrees with the $27.42 shown at the checkout page the reviewer reached.
• The JointVive review found a "Scientific references" list that cites papers on prostate health, testosterone and iodine, not joints. A percentage taken from an unrelated study means nothing for the product.

A six-question check for any percentage in an ad

1. Is it a relative or an absolute change? Ask for the starting rate and the end rate.
2. Percent of what, and how many people out of 100 or 1,000?
3. Was there a placebo or control group, or only a before-and-after comparison?
4. Did the study test this product, at this dose, in people like you?
5. Is the number the average result, or the best dose, best subgroup or best outcome?
6. Is the change big enough to matter in daily life, not just on a lab test?

Our how we review page explains how we weigh label doses and cited studies, and the heart health category shows how these checks play out on blood pressure offers.

FAQ

Is a relative risk reduction always misleading?

No. It is a valid measure, and the Cochrane Handbook notes that relative effects tend to be more stable across risk groups than absolute ones. It becomes misleading when it appears alone, without the baseline risk and the absolute difference.

What does "clinically proven to improve by 40%" actually tell me?

Very little on its own. You need to know who was measured, against what comparison, for how long, and whether the improvement was larger than in a placebo group. The FTC guidance says improvements in the treatment group alone could come from a placebo effect or other variables.

How can I turn a percentage into something I can picture?

Use whole numbers. If the ad says "50% fewer", ask how many people out of 1,000 had the problem with and without the product. Cochrane describes this "natural frequency" style as the one consumers understand best.

Should I ask my doctor about a supplement that quotes big percentages?

Yes. A doctor or pharmacist can check the product against your medicines and health conditions, and can tell you whether a promised change is realistic. This matters most for blood pressure, blood sugar and other areas where a medicine may already be involved.