A rejected ad is rarely one ad. It is a signal about how the account is being run, and Meta treats it that way.
One rejection is an annoyance: you edit the copy, resubmit, lose half a day. The cost is the delay. What changes the economics is the pattern. Repeated policy violations against the same ad account move it from automated review into restriction, and a restricted account is not a slow account, it is a stopped one. For an agency that means a client whose spend is frozen through no fault of their own, and a conversation you did not budget for.
This is a guide to the rejection reasons that actually recur, why they recur, and what a check before publishing has to look at to catch them.
The rejections that keep coming back
Claims the landing page cannot support. The single most common pattern across client accounts. The ad promises a result, a timeframe or an earnings figure, and the page behind it does not substantiate it. Reviewers compare the two. So does the law in several jurisdictions.
Personal attributes. Copy that implies the reader's health condition, financial situation, race, sexual orientation or similar. "Struggling with debt?" reads as helpful and is a policy violation, because it implies knowledge about the person seeing it. The fix is almost always to describe the product rather than the reader.
Before-and-after imagery. Banned outright for body and health, and the ban is broader than most people assume. A split image showing any transformation is risky even when the subject is a kitchen.
Engagement bait and fake interface elements. A play button drawn onto a static image, a fake cursor, a fake comment field. These are rejected on sight and are also one of the clearest signals of a low-quality advertiser.
Restricted categories handled as if they were ordinary. Credit, housing, employment, health, politics. Each has its own rules and in the US several are governed by law rather than platform policy, which is the part that catches agencies out.
What it actually costs
The ad you lost is the smallest number in this. Count instead:
- The campaign that missed its window because approval took three resubmissions.
- The creative you cannot reuse, because it was built around a claim you cannot make.
- The account strike, which is invisible until there are enough of them.
- The client conversation, which costs more trust than the ad was worth.
An agency running dozens of accounts absorbs this as background noise. It is not background noise. It is a recurring tax on velocity, and it is almost entirely avoidable.
Platform rules are only half of it
The common mistake is treating this as a Meta problem. It is not. The rules an ad has to satisfy come from two different places and only one of them is the platform.
Platform policy changes often, is enforced automatically, and differs per network. What passes on Meta can fail on TikTok and vice versa.
Law does not change often but carries consequences that dwarf a rejected ad. In the United States the FTC's Endorsement Guides govern influencer disclosure, and its rule on fake reviews and testimonials carries civil penalties. Earnings claims, health claims and negative-option billing each have their own requirements. Housing, credit and employment ads are covered by the Fair Housing Act, Regulation Z and EEOC rules respectively.
In the European Union the picture has moved recently and keeps moving. Public advertising of prescription medicines is banned. The Omnibus directive governs how a price reduction may be displayed. The green-claims ban applies from 27 September 2026, and the AI Act's deepfake labelling obligation applies from 2 August 2026. The Digital Services Act limits advertising to minors.
In the United Kingdom the CAP Code and ASA rulings apply, including a standing ban on advertising Botox to the public.
Your reviewer at Meta is not checking most of that. A regulator might.
What a useful check looks like
Reading the policy page is not a process. A check that actually prevents rejections has to run where the ad is written, not after it is submitted, and it has to look at more than the words.
The copy and the settings. Wording, targeting, media specs. This is the fast part and it should be instant, because a check that interrupts you is a check people turn off.
The landing page. Fetched from outside your own network, because a page that loads for you may not load for a reviewer. Does it resolve, is it secure, does it have a privacy policy, and does it support the claim the ad makes.
The creative itself. Keyword rules cannot see a before-and-after photo, a drawn-on play button, or a landing page whose subject does not match the ad. Reviewing images needs a model, not a word list.
A record of the decision. Who checked, what it flagged, and what they chose to do. If a client or a regulator asks what was reviewed, "we looked at it" is not an answer.
This is the shape of the check Inflowave is building into the ad builder and the post composer: a rule catalog covering ten networks and four legal systems, every rule carrying a link to the policy page or statute it comes from.
Where judgement still belongs to you
Two honest limits worth stating.
Industry-specific rules depend on knowing the industry, and most client records do not have one saved. Guessing it from the copy works, but it is a guess, and a supplement brand misread as general e-commerce will not get the FDA rules applied.
And no automated check is legal advice. It catches common, well-documented problems. For a regulated client, a human who knows the sector still reads the ad.
The short version
Rejections are a systems problem, not a copywriting problem. They recur because nothing in the workflow looks at the ad before the platform does. Move the check to the moment the ad is written, include the landing page and the image, and keep the record. The rejection rate stops being a fact of life.
