Supplement Ads That Pass Review: The Meta and Google Compliance Playbook

What gets supplement ads rejected on Meta and Google?

Three things cause most rejections: claims that a product treats, cures, or prevents a condition; copy that calls out the reader's personal health status ("Do you suffer from..."); and imagery that promises transformation, before-after frames included. Both platforms also read your landing page, so a clean ad in front of a wild funnel still fails.

TL;DR

• Nearly every supplement rejection falls into three classes: disease and treatment claims, personal-attribute call-outs, and transformation imagery.
• The framing that passes is structure-function language: supports, maintains, contributes to, backed by substantiation you can actually produce.
• EU and US operators play different claim games. The EU authorises claims per ingredient in a public register; the US allows structure-function claims under FTC substantiation rules.
• Review reads the whole funnel. Your advertorial and product page are part of the ad.
• How you handle a rejection matters almost as much as avoiding one: edit precisely, appeal only false positives, replace borderline ads.
• Compliant copy is not weak copy. Mechanism-led angles sell hard without a single banned phrase.

Three rejection classes cover almost every disapproval

Sort a hundred rejected supplement ads into piles and you end up with three. We know because sorting rejected ads is a genuine part of this job.

Disease and treatment claims. The moment copy says a product treats, cures, prevents, or heals a condition, it stops being a supplement ad and becomes a medical claim neither platform will run. "Cures insomnia." "Treats anxiety." "Heals your gut." Google's healthcare and medicines policy draws this line explicitly, and Meta draws the same one. The trap is that disease vocabulary hides inside soft verbs: fights, reverses, repairs, ends. Automated review reads those verbs exactly the way a regulator would.

Personal-attribute call-outs. Meta prohibits ads that assert or imply personal characteristics, and health status sits on that list. "Do you suffer from bloating?" implies the platform let an advertiser target the viewer's medical condition, which is the exact perception Meta's ad standards exist to prevent. The fix is grammatical, not creative: describe the condition in the third person. "Bloating after meals is one of the most common digestive complaints" carries the same hook without pointing a finger at the reader.

Before-after and transformation imagery. Split frames, "week 1 versus week 6," a zoomed shot of a body region presented as the problem. Meta's rules against unlikely results and implied transformation catch these on the visual alone; the copy never enters the decision. This class confuses operators the most, because the text was compliant and the rejection arrived anyway. The image was the claim.

As of mid-2026 most of these verdicts come from automated review within minutes of upload, and since the platforms reclassified health and wellness as a sensitive category in 2025, the screening has only tightened. Which makes the practical question: what does pass?

The framing that passes: supports, maintains, contributes to

Structure-function language describes what a nutrient does for a normal body function without naming a disease. "Supports restful sleep" instead of "cures insomnia." "Helps maintain healthy digestion" instead of "fixes bloating." It reads modest next to the claims that get rejected, and the modesty is the entire point: platforms allow statements about supporting normal function and reject promises about medical outcomes.

The two big markets formalize this differently.

US operators get the structure-function lane by statute. Supplements may claim effects on the structure or function of the body without pre-approval, but the FTC expects competent and reliable scientific evidence behind every claim you make. Build the substantiation file before the ad runs, not after somebody asks for it. An ad you cannot back is a liability with a good CTR.

EU operators work from a fixed menu. Only health claims authorised in the EU register may be used, and they attach to ingredients, never to your product as a whole. If your sleep formula delivers the specified dose of melatonin, you may say it contributes to reducing the time it takes to fall asleep, because that exact claim is authorised. Your proprietary blend, as a blend, has no authorised claims at all. Wording proximity matters too: drift far from the register phrasing and you are off the menu.

Here is what surprises people: the constraint helps the copy. When the claim has to stay modest, the persuasion moves into the mechanism, the explanation of why sleep breaks or digestion struggles in the first place. Skeptical supplement buyers convert on mechanism, not on adjectives. The compliant structure and the converting structure turn out to be the same structure.

The pass/fail claims table, with rewrites that keep the sale

This is the table we brief every copywriter with before a supplement account starts spending. The claims are generic category examples on purpose: run your own copy down the same three columns.

Read down the verdict column and the pattern shows itself. Every fail is a promise about an outcome. Every pass is a statement about a function. The rewrite never argues the product works harder; it moves the argument into the mechanism around the claim, where it belongs.

One extra filter for EU operators: run the right-hand column against the register before shipping, because a phrase that clears Meta review can still be an unauthorised claim in the EU. Platform approval is not a regulatory opinion. And no verdict is final until review has seen where the click lands.

Review reads the whole funnel, not just the ad

Both platforms crawl the destination behind the ad, at review time and again while the ad runs. A compliant ad pointing at a landing page full of disease claims fails as a unit, and repeated funnel violations mark the account, not just the ad.

For supplement brands this matters most on the pre-sell layer. The advertorial is where the storytelling lives, which makes it the highest-risk surface in the chain and the one operators most often forget review can see. The tempting move is to keep the ad clean and let the advertorial say what the ad could not. That is not a compliance architecture. That is a delayed rejection, because the page is held to the ad's standard.

What works is one line held across the whole chain. Every mechanism claim gets a substantiation note before the copy is written, the advertorial gives the mechanism room to breathe rather than giving claims a place to hide, and the product page matches both. It is the same discipline our advertorial team builds into every pre-sell page: the long format exists to explain more, not to claim more. Check the forgotten surfaces while you are at it. Quiz steps, upsell pages, and email capture screens get read too.

The account health playbook when a rejection lands

Even clean accounts catch rejections. Automated review has false positives, health vocabulary is a minefield of near-misses, and a real testing program ships a lot of ads. The rejection itself is an event; what you do next decides whether it stays one, because platforms track how an account behaves after disapprovals.

• Find the actual trigger first. The stated reason is broad; the trigger is usually one phrase or one frame. Read the ad like a reviewer: verbs first, imagery second, landing page third.
• Edit beats re-upload when the trigger is visible. Fix the flagged element and resubmit the same ad. In our experience this is the cleanest path back through review, and it preserves the ad's history.
• Appeal only genuine false positives. Compliant ads get caught on vocabulary regularly, and those appeals get overturned often enough to justify the wait. Appealing an ad that really does cross a line burns days and stacks another mark.
• Replace borderline ads instead of fighting for them. If you have to squint to argue the ad complies, write the compliant version and ship it as new creative. A new ad beats an appeal most weeks: faster back to spend, nothing lost but pride.
• Never carpet re-upload identical rejected creative. Submitting the same rejected ad unchanged, hoping a different reviewer waves it through, reads as evasion. Accounts that do this drift toward restrictions, and a restricted account is the most expensive outcome in the category.

Then measure it. We track ad approval rate in supplement accounts the way we track CPA, because a testing program that loses a third of its creative to review is not testing at the speed the account is paying for. It is a standing item in how we run Meta ads for regulated categories.

Compliant angles that still convert

Here is the part most compliance advice misses: written properly, the compliant ad usually converts better. A modest claim wrapped in a sharp mechanism reads as credible to a buyer who has scrolled past a thousand miracle promises. The overclaim does not just risk the account; it triggers the exact skepticism that kills supplement conversion in the first place.

So the build order we run is fixed: claims sheet per ingredient, then mechanism, then hook pool, then creative volume, with ad, advertorial, and product page holding the same line. We do this daily for supplement brands in both the EU and US, on both platforms. The usual caveat, and it is a real one: this is marketing operations guidance from people who run these accounts, not legal advice, so keep regulatory counsel close for labels and market-specific rules.

If review keeps eating your testing velocity, the fix starts upstream of the ad account. Our playbook for scaling supplement and nutrition brands covers the category economics, the compliance architecture, and where the 2026 openings are. Rejected ads are not the cost of selling supplements. They are the cost of writing for the category carelessly, and that one is optional.