“Depression counseling” passes. “Depression getting you down?” fails. Same subject, opposite outcomes — and three tests that catch it before review.
Question at the reader · the word “other” · “you” + attribute · names count · the landing page is reviewed
Ad rejected for “personal attributes”. The instinct is to hunt for the forbidden topic: can I not mention health? debt? age?
That question is wrong, which is why the fix usually fails two or three times. There is no forbidden topic in this policy. There is a forbidden CONSTRUCTION.
The same subject, with the same keywords, passes or fails depending on how the sentence is built. Once you see the mechanism, you can check any creative in ten seconds — no guessing, no waiting for review.
You may NAME the attribute. You may not AIM it at the reader.
The policy says ads must not “assert or imply” personal attributes of the viewer. Imply is the word carrying all the weight: you do not have to claim the person has a condition — the sentence only has to make sense if they do.
Notice the approved side describes the PRODUCT. The rejected side describes the READER. That is literally the policy guidance: ads should focus on the benefit of what is being advertised, not on who is on the other side of the screen.
This is the most common trap, because a question is the most natural copy device there is. “Are you self-employed?”, “Is your business in debt?”, “Over 50?” — each one qualifies the audience, and each one implies the advertiser knows something about the reader.
Rule: if the sentence is a question whose answer reveals an attribute, rewrite it as a description of the offer. Targeting already does the job of finding the audience — the creative does not need to screen people out loud.
This is the subtlest of the three, and the one that surprises people who already know the policy. Two sentences identical in subject, separated by a single word, land on opposite sides.
The reason is pure grammar: “other X” only makes sense if the listener is an X. The word carries the claim about the reader without anyone having written the claim. That is exactly the kind of implication the policy forbids.
Rule: search for “other”, “others”, “too”, “like you” and “fellow” in front of a group. They are all the same trap wearing different clothes.
Here lives the industry's biggest misreading: plenty of people believe Meta bans the word “you”. It does not. The policy explicitly allows second-person language — what it blocks is “you” glued to an attribute.
The skincare pair is worth sitting with: “Our new creams fight wrinkles like never before” passes, and “Ready to upgrade your skin to look younger?” fails. Same product. What changed was the subject of the sentence.
Rule: keep the “you” — it converts. Just don't let it touch age, health, money, belief, origin or condition. Move the subject of the sentence to the product and the line survives.
The covered attributes: race, ethnicity, religion, beliefs, age, sexual orientation or practices, gender identity, disability, physical or mental health, vulnerable financial status, voting status, trade union membership and criminal record. So far, predictable.
The name one surprises people because personalization is sold as best practice in nearly every other channel. In Meta creative it is not: implying the advertiser knows the viewer's name sits in the same violation family as implying an illness.
The policy also forbids implying knowledge about the person's FAMILY — financial or medical. “Get your spouse treated today” is cited as a rejected example.
Two lines in the policy go unnoticed and outweigh any copy tweak, because they change how big the damage gets when something goes wrong.
Meta states plainly that it monitors advertiser behavior, not just individual creatives. For anyone running several accounts from the same personal profile, that is systemic risk: repeated rejections on one client become a problem for all the others.
There is a third, quieter line: low-quality ads that don't violate any policy can still have their delivery impacted. They aren't rejected, they raise no warning, they just underperform — and you spend weeks blaming the creative team.
Rule: treat the landing page as part of the creative and account structure as part of the risk. Business Manager partnerships exist precisely for this — the relationship becomes company to company, instead of one individual sitting inside dozens of accounts.
Is it a question whose answer reveals an attribute of the reader? Turn it into a description of the offer.
Does it say “other”, “too”, “fellow”, “like you” in front of a group? Delete the word — the sentence almost always survives.
Does “you/your” touch age, health, money, belief, origin or condition? Move the subject to the product.
Personal names and financial status count as attributes. Name-personalized creative is a violation, even though it's best practice elsewhere.
The landing page is reviewed, and the penalty reaches the whole Business Account — ad account, Page and user account.
Describe the product, never the reader. Almost every personal-attributes rejection is a sentence that moved the subject — and almost every fix is moving it back.