Your Ad Account Doesn’t Have a Copy Problem. It Has an Evidence Problem.
Why Google and Meta suspend supplement ad accounts, and what separates the brands that get pulled from the ones that don't.
The copy gets rewritten. The landing page gets audited. A compliance consultant gets paid to sweep the site for language that the platform might flag.
Four weeks and a few rewrites later, the ad gets pulled again. Same problem, different wording. The problem is not in the copy itself. It is in what sits behind the copy.
Google’s Healthcare and Medicines policy requires that health claims are factual and verifiable. Meta’s ad standards require that claims are truthful, not misleading, and supported by evidence. The FTC requires “competent and reliable scientific evidence” before you make any health benefit claim.
The words are the symptom. The evidence, or the absence of it, is the cause.
Here is what I see founders getting wrong, and what the ones keeping their accounts alive are doing differently.
What we’ll cover today:
What Google and Meta are actually checking for
Why rewriting your copy is not the solution
How to tell what evidence holds up — and what the FTC says about it
How to unlock claim language your competitors can’t copy
How to sequence research before ads, and why it matters
A note for UK and EU brands — why the rules are different
P.S. Save this article. The next time an ad gets flagged, you’ll want sections 3 and 4 open. And if someone on your team handles ad accounts or compliance, send it to them. The common mistakes section alone is worth their time.
1. Account survival lives in evidence, not wording
Both Google and Meta use automated systems to handle the first pass of ad review.
These systems scan for before-and-after framing. Superlatives attached to health outcomes. Language that connects your product to a specific condition. Body function claims about “boosting metabolism” or “balancing hormones.” Implied diagnostic language.
If your copy trips one of those patterns, the ad gets blocked.
But here is the part most founders miss. If you appeal that decision, or if the platform escalates your account for deeper review, a human gets involved. Google has confirmed it uses a combination of AI and human evaluation, where AI models either block policy-violating content outright, or flag it for trained reviewers when a more nuanced determination is required.
That human reviewer is looking for something behind the patterns. Most founders only prepare for the automated check. That is why the account keeps getting pulled, they pass the first check and fail the second.
Key Point. The automated system decides whether your ad gets flagged. The appeal or escalation process decides whether your account survives. The first is a copy question. The second is an evidence question.
2. Softer copy doesn’t make your account safer
Most brands respond to a suspension by softening the language.
“Reduces bloating” becomes “supports digestive comfort.” “Shown to improve sleep” becomes “helps you unwind.”
You strip out the superlatives, you remove the before-and-afters, you swap condition words for function words.
This works for a few weeks, then a different line gets flagged. You are back in the same position, except now the appeal process is harder. The reason it keeps happening is that you have made the copy more vague, but the underlying question has not changed.
The platform is still asking: can you prove what you are saying?
Vague function language lowers your automated flag rate, but it gives you nothing to show in an appeal or manual review. The trade-off is worse than it looks. By using the same language every other brand in your category can use, you have softened yourself into the middle of the market.
No differentiation and no protection.
3. Product-level evidence is what holds up and the FTC agrees
When your account lands in manual review, the reviewer is looking for documentation that maps to your claims. Not every kind of evidence carries the same weight.
Ingredient studies: generic research on an ingredient you happen to include in your formula carry the least weight. They do not speak to your product, your dose, or your consumer.
White papers and internal data carry slightly more, but reviewers treat them as marketing documents, not evidence.
Published, peer-reviewed studies on your actual formula carry the most. This is the strongest kind of evidence you can bring to a review, and the hardest for anyone to challenge on relevance.
This is not just my experience working with brands through the review process. The FTC’s 2022 Health Products Compliance Guidance says it directly. The FTC states that a “clinically tested ingredient” claim on a product not only implies that the ingredient has a benefit, but also that the product containing the ingredient confers that benefit. Because the product may contain other ingredients, the FTC says an advertiser should consult with a qualified expert to determine whether experts in the field would generally require a clinical test of the product itself, rather than the isolated ingredient, to confirm the benefit.
Federal courts have gone further. In FTC v. Wellness Support Network, the court accepted expert testimony that RCTs for dietary supplements should be conducted on the exact dosage and formulation rather than on individual ingredients, because “there may be interactions between the ingredients that affect their physiological actions.”
This is what I call product-level evidence. Evidence you have generated yourself:
On the exact product you are selling
At the dose you are selling it at
To the consumer you are selling it to
Platform reviewers and regulators are looking for the match. No match, no proof, no ad.
Key Point. If your strongest documentation is a study on a single ingredient used at a different dose than yours, the FTC’s own guidance says that may not be sufficient. Ingredient studies can support your position, but only if the dose, form, and population match your product. If they don’t, the reviewer has no reason to connect them to your claim. Plan accordingly.
4. Product-level evidence unlocks claim language no competitor can use
Here is the part most founders miss.
Product-level evidence does two things at once.
It protects your ad account
It opens up claim language nobody else in your category can use
Without it, the best you can say is “supports X.” Every competitor can say the same.
With it, you can say: “in a 12-week study on our formula, participants saw a Y% improvement in X.”
That claim is specific, it is verifiable, and it maps to documentation the platform can check. You go from fighting the platform on every ad to having a set of claims the platform has no reason to question.
Important: this specific claim language applies to US-market advertising under FTC structure/function rules. If you are a UK or EU brand, the claim landscape is different — see Section 6 below.
5. Research before ads, not after
There is a way to change the order of operations.
Most brands launch ads, get suspended, rewrite the copy, get suspended again, and only then start asking about research. That sequence is backwards. The research should come first. A small, properly designed study on your formula, run before you scale paid media, with the ad account built on top of it.
In practice, that means running a pilot study ahead of your next big ad push.
Remote, placebo-controlled, randomised, double-blind. It runs in roughly six months from protocol to data, rather than the 18–24 months a traditional academic trial takes once you factor in ethics review, funding cycles, and publication. It gives you product-level numbers you can cite in your ads, and documentation that holds up in a manual review when it comes.
The pilot gives you proof behind your claims. It unlocks differentiated language your competitors cannot copy. And it empowers the study you will design for publication later.
6. A note for UK and EU brands
Everything above applies cleanly to the US market, where the FTC operates on a substantiation model: prove what you say, and you can say it.
The UK and EU work differently, and the difference matters.
Under the Nutrition and Health Claims Regulation (retained in UK law as the GB NHC Register), you can only use health claims that have been pre-authorised and listed on the Register. It does not matter how strong your evidence is — if the claim is not on the Register, you cannot make it.
You cannot say “clinically proven.” You cannot cite specific study results in your ads. The ASA has repeatedly upheld complaints against supplement brands using those phrases, even when clinical evidence existed. The issue is not whether the proof is real. It is that the claim type itself is not permitted.
Claims are authorised at the nutrient level, not the product level. “Zinc contributes to normal cognitive function” is permitted. “Our formula improved cognitive function by 15%” is not — regardless of the evidence behind it. Even small wording changes can breach the rules. “Contributes to” cannot become “improves” or “increases.” The ASA has ruled that dropping the words “contributes to” from an authorised claim constitutes an exaggeration.
To make an authorised claim, you need the right amount of the nutrient in the product. That is it. If your formula contains sufficient zinc to meet the conditions of use, you can make the authorised zinc claim. You do not need an RCT for that. So if the value of product-level evidence in the UK were only about compliance, it would be a hard sell.
But compliance is not where the value sits. The real value is everywhere else.
The advertising rules lock down your ads. They do not lock down your brand.
The GB Register governs paid advertising and product labels. It does not govern what goes on your website, what you discuss on a podcast, what you present at a conference, what you put in a pitch deck to a retailer, or what a journalist writes about your product after reading your published research.
And increasingly, that is where supplement marketing actually happens. The brands that are winning in the UK market right now are not winning because of what they can say in a Meta ad. They are winning because of the credibility they carry everywhere else.
A study on your actual formula, not borrowed ingredient research, but your product, your dose, your consumer, gives you material that works across every channel the advertising rules do not touch:
Your website. You can present your study methodology, your results, and your data in full. No character limits. No automated review. Consumers who care about evidence, and that segment is growing fast and they will find it.
Podcasts and editorial. A founder who can say “we ran a double-blind, placebo-controlled trial on our formula and here is what we found” gets invited back. A founder who can only say “our product contains ingredients that have been studied” does not. The study gives you a story worth telling and a reason for someone else to tell it.
Conferences and trade shows. If you sell B2B, to retailers, distributors, or practitioners, the study becomes your most valuable sales asset. It answers the question every buyer is really asking: how do I know this product works?
PR and organic coverage. A published study is a newsworthy event. It creates a reason for industry press, health journalists, and influencers to cover your product without you paying for the placement.
This is what I call scientific brand alignment. It is the difference between “we sell a product with ingredients that have been studied” and “we are a company that studies our product.” That distinction matters to retailers, to investors, to podcast hosts deciding who to feature, and increasingly to consumers who can tell the difference.
And there is one more move. Article 13(5) of the retained regulation allows for the authorisation of health claims based on newly developed scientific evidence, including requests for the protection of proprietary data. If your RCT generates genuinely novel evidence, you can apply for a new authorised claim that only your product can use. That is the UK equivalent of “claim language your competitors cannot copy.” It is a longer play, but it is the endgame.
Most common mistakes I see
Appealing with the same claims and different wording. Google evaluates whether you have genuinely resolved the underlying issue, not just acknowledged it. Saying “we fixed our landing pages” means nothing. Explaining that you removed specific unsubstantiated claims and added substantiation documentation gives the reviewer something to work with.
Creating a new account on the same domain after a rejection. Google treats this as attempting to circumvent their system, an egregious violation that triggers immediate suspension without prior warning. Do not do it.
Leaning on a supplier’s ingredient study because “there is research behind it.” The FTC’s own guidance says studies on individual ingredients may not be sufficient to substantiate a claim about a combination product, because the ingredients together may affect the body differently than they do individually. Reviewers can tell the difference between ingredient research and product research, and so can your consumers.
Letting your compliance person rewrite claims without checking what evidence your research lead has generated. If they are not talking to each other, the claims and the evidence will not match. That is where the gap opens up.
Key Point. If your account has already been suspended, start with Section 3. Get your evidence sorted before you appeal again. If your account is still live but your strongest proof is an ingredient study at a different dose, Section 5 is the move. If you are a UK brand, Section 6 is essential reading.
The Bottom Line
Your ad account does not have a copy problem. It has an evidence problem. Fix the evidence, and the copy stops being the thing that gets you pulled.
— Nathan
P.S.
If this problem sounds familiar, at StudySetGo we run pilot studies on your actual formula, remote, placebo-controlled, double-blind, gold-standard.
This allows you to have product-level evidence behind your claims.
The pilot runs in roughly six months from protocol to data, costs a fraction of a university trial or CRO trial, and the documentation it produces is the same documentation a manual reviewer is looking for.
For US brands, that means stronger claim language your competitors cannot copy. For UK and EU brands, it means scientific brand alignment, the study that powers your website, your podcast appearances, your B2B sales conversations, your PR, and eventually a proprietary health claim application on the GB Register.
If you want to map out the fastest path to evidence for your product, book a no-obligation conversation with me.
If this is your first time reading, this newsletter covers how supplement and health brands use science to grow — and stay compliant while doing it. I send it every week on Tuesday. Subscribe here so you don’t miss the next one!