
Every year Meta updates something, and most of the time you can safely ignore it — a few tweaks to the wording, some cosmetic changes to policy language. But 2026 is a different situation. This year the platform pushed through 47 documented advertising policy changes, and a good chunk of them don't just adjust the details — they change the underlying logic: how ads get reviewed, who's responsible for claims made in creatives, which products now require a special category just to run. If your campaigns are set up the same way they were a year ago, some of them are already out of compliance — they just haven't been banned yet.
Below is a before/after breakdown of the key changes, the three shifts that matter most for affiliates specifically, and a checklist for a quick self-audit.
Before vs. After: The Key Changes
Meta Ads 2026 — Policy Changes
|
Area |
Before 2026 |
After 2026 |
Impact |
|
Ad review |
Reactive — user complains, then review |
AI scans every ad before the first impression is served |
No more "test and see." Rejections happen before any spend. |
|
AI creatives |
No disclosure required |
Mandatory "AI-generated" label on all AI visuals and audio |
14% of all rejections now come from undisclosed AI content |
|
HEC |
Advertiser selects category manually |
AI auto-classifies ads with housing/credit/employment imagery |
Ads previously running unrestricted may be flagged retroactively |
|
BNPL |
No special category required |
Classified under Credit — full HEC restrictions apply |
Klarna, Afterpay, Affirm need Special Ad Category + verification |
|
UGC |
Could run as an organic-looking ad |
Must use Partnership Ads format — any compensation triggers this |
Without the format: Deceptive Practice violation, account health hit |
|
Liability |
Creator's responsibility |
Brand/affiliate equally liable for claims in boosted Partnership Ads |
"10 min approval" in a creator video is your compliance exposure |
|
Review scope |
Text and image reviewed separately |
Text + image + video + audio + landing page reviewed as one unit |
Landing page mismatches now cause ad rejection |
|
Attributes |
Only direct statements banned |
Indirect implications also banned ("if you struggle with...") |
Health and beauty rejection rates spiked 34% |
|
Crypto |
Single authorization gate |
Three-tier system — no verified badge means all ads auto-rejected |
Requires regulatory license, custody insurance, audit certification |
|
Verification |
Required in 12 countries |
Required in 38 countries |
Verify advertiser status in every new target market |
|
Audiences |
Financial indicators (income, net worth) allowed |
Banned as of September 2025 |
Rebuild any audiences built on financial signal data |
|
Targeting |
Detailed interests treated as hard constraints |
Treated as suggestions — Advantage+ expands beyond them |
Creative and offer now do the targeting work |
|
EU fees |
Standard pricing |
Fee added on top of budget, per country, from July 1 |
UK +2% · FR/IT/ES +3% · AT/TR +5% |
|
Recovery |
Informal appeal process |
4 steps: training → action plan → quiz → 30-day manual review |
Getting banned is significantly more expensive in time |
The Three Changes That Hit Affiliates Hardest
1. Your landing page is now part of the ad review. Meta's review system no longer looks at your ad in isolation — it reads the first fold of your landing page as part of the same submission. That means an ad copy that passes every check can still get rejected because of a claim on your landing page that wouldn't have triggered anything on its own. Before launching anything, review the ad and the landing page together as if they were a single piece of copy.
2. Partnership Ads are mandatory, not a format you opt into. Any creator who received money, a gifted product, or an affiliate commission is covered by this rule, regardless of how informal the arrangement was. If they make a claim in that content — say, "fastest approval I've ever seen" — and you boost it, that claim becomes your compliance exposure, not theirs. The fix is either briefing creators on what they can and can't say, or simply not boosting content you haven't reviewed.
3. Fintech and BNPL get auto-classified whether you declare them or not. Meta's image-scanning system now detects loan calculators, credit card mockups, and BNPL checkout flows in your creatives and automatically applies Special Ad Category restrictions — even if you never selected the category yourself. This happens retroactively, so campaigns that have been running fine can suddenly find themselves restricted. If you're in any of these verticals, check your active campaigns now.
Compliance Checklist
Creatives
-
Audit all active ads for AI-generated content — add disclosure label
-
Verify no UGC or creator content is running outside Partnership Ads format
Copy and landing pages
-
Remove conditional/empathy phrasing ("if you struggle with...", "we know how hard...")
-
Align landing page first fold with ad copy — no claims that don't appear in the ad
Account structure
-
Check HEC classification — especially BNPL and lending products
-
Confirm advertiser verification in all target markets (38 countries now required)
-
Remove Custom Audiences built from income, net worth, or creditworthiness data
EU campaigns
-
Recalculate ROAS benchmarks with location fees factored in (UK +2%, FR/IT/ES +3%, AT/TR +5%)
What Didn't Change
Worth separating from the compliance picture: Advantage+ automation, the Andromeda and GEM delivery engines, and Meta's generative creative tools are all platform infrastructure updates — they shift how your campaigns perform, but they don't create compliance risk on their own. The policy changes in the table above are where actual account bans come from, and that's where your attention should go first.



