Eleven things operators believe, and the source text that contradicts each

Most of what circulates about ad compliance is not wrong invented — it is wrong remembered. A rule got summarised, the summary got repeated, and the version that survives is the one that was easiest to say. The instruments themselves are public and mostly short, and the gap between what they say and what the channel believes is where the expensive afternoons come from.

So this page is eleven of those beliefs. Each one is followed by a single line from the source that contradicts it, and then by where the whole question is worked out. No interpretation of ours sits between the belief and the quotation — that is the point of the format. If you only read the quotations you will still be ahead.

Eleven beliefs, and the line that answers each

“It was approved, so it’s allowed.”

“Ads remain subject to review and re-review at all times, and may be rejected or restricted for violation of our policies at any time. It is your responsibility to understand and comply with our policies.”

Approval is a moment, not a permission. The sentence is Meta’s own, in the Advertising Standards, and it means a creative running profitably today can be refused tomorrow with nothing having changed on your side — and that an edit resubmits the ad as a new one. Worked out in the guide on the three outcomes, which also covers the one that arrives with no notice at all.

“The rejection tells me which line to fix.”

“Unacceptable business practices · Discriminatory practices · Social issues · Misrepresentation”

Those are the labels a rejection arrives as. Not one of them is a sentence in your ad, and none of them names the phrase that triggered it. The label comes from a reader whose question is not the regulator’s: the platform asks what your copy promises the person reading it, the law asks what it claims about the product. Knowing which of the two you hit is the difference between a ten-minute fix and an afternoon.

“I only get paid if someone buys, so there’s nothing to disclose.”

“the possibility of being paid, of winning a prize, or of appearing on television or in other media promotions”

That phrase is inside the definition of a material connection in 16 CFR §255.5(a). A commission that has not arrived and may never arrive is a connection; so is a free product, “regardless of whether the advertiser requires an endorsement in return”. The duty does not wait for the money.

“The paid-partnership toggle is my disclosure.”

““Don’t assume that a platform’s disclosure tool is good enough, but consider using it in addition to your own, good disclosure.” — FTC, Disclosures 101 for Social Media Influencers”

The Commission wrote out the case where relying on the platform’s own tool fails, and what failed was not the label’s wording — it was size, contrast, competing text and five seconds on screen, all decided by the creative built around the label rather than by the platform that supplied it. Use the toggle if you like it, then put your own line in the creative.

“#ad works everywhere.”

“The disclosure must be in the same language as the ad.”

Example 11 in §255.0(g) holds that a translated ad carrying an untranslated disclosure is inadequate — the claim reaches the reader and the qualification does not. Germany goes further and expects the German words: the media authorities warn against the English short forms because they are used inconsistently and easily missed. The label is not a fixed token to be copied across markets.

“Vague copy is safe copy.”

“Reference to general, non-specific benefits of the nutrient or food for overall good health or health-related well-being may only be made if accompanied by a specific health claim included in the lists provided for in Article 13 or 14.”

This is the one that runs against everybody’s instinct, and European law reverses it outright: a soft wellbeing line is not unregulated, it is conditionally lawful, and the condition is that a specific authorised claim travels next to it. Claims judged too general even to assess were not waved through — they are recorded in the register’s non-authorised list.

“If I can prove it, I can say it.”

“A health claim may be used only if it is on the authorised register, in substantially the wording that was authorised, and within its conditions of use.”

True in the United States, where the question is whether you can substantiate what consumers take away. Not true in the European Union or Great Britain, where the question is whether the wording is on a closed list. Evidence gets a claim ONTO that list, through an authorisation procedure; it does not license a sentence that is not on it yet. The two models fail in different places, which is why a funnel that runs in both needs both checks.

“‘Results not typical’ covers an unusual outcome.”

“The 2023 revision removed that cure: the advertiser now needs substantiation that consumers generally achieve the depicted result.”

It was the standard patch, and it stopped working. Worse for anyone commissioning creator content: since 21 October 2024, 16 CFR Part 465 makes compensation conditioned on a positive sentiment a violation in its own right — which is a clause sitting inside a great many existing briefs, phrased as “must say it works”.

“A countdown is just marketing.”

“Three of the four markets we cover complete the breach on falsity alone, and in one of them it is a criminal offence.”

False urgency is unusual in this field because it needs no interpretation. Almost every advertising breach requires somebody to argue about an impression; this one requires a page reload. If the timer restarts at fifty-nine minutes, you have produced the evidence — and so can a competitor, a consumer group, or a regulator with a browser.

“The American rule was struck down, so the practice is legal again.”

“FTC Act §5 on deceptive acts and practices, and ROSCA for online negative-option transactions. Plus the original 1973 Negative Option Rule, which came back into effect when its replacement was vacated.”

A vacated rule is not a permission. The Negative Option Rule was set aside on procedural grounds, which removed a specific instrument and left the general prohibitions standing — and it revived the older rule the replacement had superseded. Meanwhile the British subscription regime is written but not commenced and the European provision has applied unchanged since 2005, so the same funnel is answerable to three clocks at once.

“An ad is judged on what it says.”

“Any claim referring to the RATE or AMOUNT of weight loss, under Article 12(b) of Reg 1924/2006. Truth, evidence and disclaimers do not enter into it.”

It is judged on what it conveys, and the instruments say so in their own words. Article 12(b) bans the reference itself rather than the untruth. And German supplement law reaches what is merely implied — including through “Darstellungen”, which is how a picture with no text at all becomes the claim you are answering for.

What the eleven have in common

Ten of these eleven beliefs are true of something. Approval does mean something. Evidence does decide the American question. “Results not typical” did work. Each one is a rule that was correct in one market, or in one year, or against one of the two readers, and then travelled — losing the condition that made it true on the way. That is the failure mode worth carrying away, more than any single item above: the belief is rarely invented, it is usually a true thing with its scope filed off. The guides below are each one belief, taken back to the instrument it came from.