Restricted Products Advertising: What Meta, Google, TikTok and Snap Actually Gate

There is a category of advertiser who has concluded the platforms are simply against them. Ads rejected, appeals ignored, accounts disabled, a replacement account disabled faster than the first. Somewhere in that sequence they start searching for workarounds, because the honest route appears to be closed.

In most cases it is not closed. It is gated, and the gate is documented. Google, Meta, TikTok and Snap all publish what they call restricted categories: products that can be advertised, but only after a certification, an authorisation, a licence submission or a pre-approval. The list of genuinely prohibited products is much shorter than the list of restricted ones, and a very large share of the accounts that die in these verticals die because nobody read the gate requirements before spending.

This is the reference page for that problem. It covers what each platform restricts, how the gates differ, the enforcement mechanics that actually end accounts, and what changes when the market is India rather than the United States. Every claim here is dated, because policy in this space moves every quarter.

Restricted is not the same as prohibited

The distinction decides whether you have a compliance problem or a business model problem, and it is worth being precise about it.

Prohibited means no version of the ad is permitted, no application exists, and no account structure changes the answer. Counterfeit goods, most weapons categories, recreational drugs, and in India since August 2025, online money games.

Restricted means the platform will run the ad once you clear a specific gate. Gambling, crypto and financial services, healthcare and pharmacy, dating and companionship, alcohol, lending, supplements, political and social issues. Each has its own gate, and the gate is per platform and usually per country.

If your product is restricted, the work is administrative: assemble the evidence, apply, wait, build the landing page to the standard the policy asks for. If your product is prohibited in the market you are targeting, no agency and no account structure fixes that, and anyone telling you otherwise is selling you something that will end badly.

How each platform gates a restricted category

Google: certification, per country, per category

Google runs the most formal system of the four. Gambling, crypto, healthcare and pharmacy, addiction services, financial services and, since 2026, dating and companionship all require certification before ads serve.

Three structural details catch people out. Certification is granted per country, so targeting five markets means five applications. Some certifications are mutually exclusive on a single account: an account can be certified for real money gambling or for social casino, not both (Google Ads gambling and games policy). And certification is applied for at the child account level rather than the manager level, which means account structure has to be settled before the applications go in, not after.

Google also now checks the domain itself. Sites on free platforms, subdomains whose root belongs to a third-party host, and second-level domains the advertiser does not own and operate are ineligible under the gambling policy, extending across every category under that policy from 14 September 2026.

Meta: written permission and authorisation

Meta uses a permission model rather than a certification one. Gambling requires authorisation requested through the Authorizations and Verifications tab in Business Suite, with evidence that the activity is licensed by the relevant regulator (Meta Transparency Center). Dating requires prior written permission through an application form, after which ads may only target people over 18 (Meta Transparency Center).

Meta's authorisation scope is wider than Google's in one important way: for gambling it reaches affiliate and aggregator landing pages even where no gambling happens on the page itself. An affiliate who assumed they were outside the policy because their site only compares operators is inside it.

TikTok: pre-approval and a rolling strike window

TikTok restricts alcohol, dating applications, financial services and over-the-counter medication with age gating and geographic limits, and requires pre-approval for financial products, health supplements and political advertising before campaigns go live (TikTok regulated commercial activities).

The enforcement model is different from the other three. TikTok runs an escalation framework on a rolling 12 month window, so violation counts reset after 12 clean months. Permanent bans, however, are not reversible. That combination rewards advertisers who slow down after a first strike and punishes the ones who push through it.

Snap: licence proof per jurisdiction

Snap asks for the paperwork directly. Advertisers in gaming and gambling must comply with all applicable licensing or registration requirements and must provide Snap with proof of current licence or registration from the regulator in each country or jurisdiction where they want to promote (Snap ad category requirements). Snap names the United States, France, the United Kingdom, Canada, Australia, Germany and Spain among markets where certain online gambling ads may be permitted.

Snap also banned political and issue-based advertising in 2024 and the ban still stands, which is broader than any of the other three. Advocacy content that would clear Meta's special ad category process has nowhere to go on Snap.

The same product, a different answer in every market

This is the part that costs the most money, because it is invisible until you scale. A category is not restricted or unrestricted globally. It sits in one of three states per market: open, gated behind a certification or licence, or closed with no application route at all. A product that is a licensed mainstream advertising category in one country can be a criminal offence to promote in the next.

Five examples that cover most of the confusion:

Gambling has the widest spread of any category. Licensed and mainstream in the US and UK. Prohibited outright in Italy under the Dignity Decree and in India under the 2025 Act. Restricted to a 01:00 to 05:00 television window in Spain, with celebrity endorsement banned. Untargeted mass advertising prohibited in the Netherlands. A 21:00 to 06:00 window for virtual slots and poker in Germany. Near-total ban in Belgium. And in Australia from 1 January 2027, no athletes, celebrities or influencers at all. Detail per region: Europe and the UK, Brazil and LATAM, Australia, United States, India.

Prescription pharma is a two-country freedom. The United States and New Zealand are the only high-income countries permitting direct-to-consumer advertising of prescription medicines. The UK, EU, Canada, Australia and most of the world prohibit it. A US telehealth or pharma funnel that performs well is built on an advertising permission that exists almost nowhere else, so extending it internationally is a different product marketing strategy rather than a localisation project.

Crypto is licence-gated per jurisdiction. FinCEN MSB registration plus state money transmitter licensing in the US, FCA registration in the UK where advertising outside the financial promotions regime is a criminal offence, and MiCA authorisation across the EU, Iceland, Liechtenstein and Norway.

Dating is closed across much of South Asia and MENA on Google. India is not on the ineligible list, but Pakistan, Sri Lanka, Nepal and most of the Gulf are, which closes the natural regional expansion path for a South Asian dating product.

TikTok has no India market at all. The app has been banned since June 2020 and remains banned in 2026. Indian businesses reach TikTok only through a foreign entity targeting foreign markets, which is an entity and banking question before it is a media question.

The full mapping of categories against markets is in the restricted ads country matrix.

The EU layer that applies whatever your category

Three rules apply across the bloc regardless of your vertical, and they catch advertisers who have read the platform policy and nothing else. The Digital Services Act bans profiling-based advertising to minors, and that ban cannot be overridden by parental consent. It prohibits profiling-based advertising using special categories of data including religion, ethnicity, health, sexual orientation and race. And it requires ads to be labelled with who placed them and why, with very large platforms maintaining public ad repositories, which means European creative is continuously inspectable by regulators, competitors and complainants.

GDPR and ePrivacy consent sit alongside that. Advertisers who move a US offer into Europe and watch conversions collapse are usually looking at a consent management problem rather than a creative one.

Chain liability is the trend to watch

India, Brazil and the UK have all moved obligations beyond the operator. India's 2025 Act makes advertising an independent offence from offering, so agencies, publishers and creators are inside it. Brazil's July 2026 ordinances apply consumer protection rules across the entire commercial communications chain. The UK CAP Code extension of 1 September reaches non-paid owned social by licensed operators even without a UK-registered company address.

If you are an agency, an affiliate or a creator, reading only operator-facing rules now reads the wrong half of the regulation.

The enforcement mechanics that actually end accounts

Reading the category policy is the easy half. The rules that do the damage sit in the enforcement documentation, and they are mostly about time limits and irreversibility.

• Meta, 180 days. An ad account disabled for 180 days cannot be reinstated. The clock starts on the disable, not on the day you notice.
• Google, three verification attempts. Advertiser verification is a prerequisite to appealing a suspension, and if identity cannot be verified after three attempts, no appeal is permitted at all. That outcome is permanent.
• Google, manager account contagion. From 23 March 2026, manager accounts carrying a significant volume of revoked gambling certificates lose the ability to apply for new certificates, with no published threshold and no stated reinstatement path. One client can cost an agency the category.
• Circumventing systems. Cloaking, multiple account abuse and false information during verification are enforced by suspension on detection, without the warning sequence that applies to ordinary policy violations (Google circumventing systems policy).
• TikTok, permanent bans. Strikes expire on a rolling 12 month window. Permanent bans do not.

Notice what these have in common. None of them are about the ad copy. They are about identity, account structure, and the speed at which you react. That is why account structure is a compliance decision in restricted categories and a convenience decision everywhere else.

What a compliant restricted-vertical setup actually looks like

The pattern that survives is boring and slow to start.

1. Classify before you spend. Work out what your product is called under each platform policy, in each country. Some products are three different categories on four platforms.
2. Assemble the evidence pack first. Licences, registrations, entity documents, proof of domain ownership. Most certification rejections are incomplete applications rather than rejected businesses.
3. Own the domain properly. Not a free subdomain, not a host-owned second-level domain. Multiple platforms now check this and it cannot be fixed after an application is in flight.
4. Build compliance into the landing page. Age gate, risk or responsible-play disclosure, licence display, terms, refund policy, and a claim set you can evidence. The landing page fails review more often than the creative does.
5. Separate product lines into separate accounts. Sometimes the policy demands it. Always it contains the blast radius.
6. Log the policy text on launch day. When you appeal in four months, the useful artefact is what the policy said when you built the campaign.
7. Re-read the policies quarterly. Google made three separate gambling policy changes with effective dates in 2026 alone.

Where the black hat search terms come from

A lot of people arrive at this topic through phrases like black hat performance marketing, grey hat media buying or dark area advertising. Usually they are not looking for fraud. They are looking for anyone at all who will take on their category, having concluded that no legitimate agency will.

The techniques those phrases actually describe, cloaking, farmed accounts, rented business managers, are not a shortcut around the gates on this page. They are the fastest available route to the permanent, unappealable outcomes listed above. We have written separately about what those terms mean and what each technique actually costs.

Category guides

Deeper detail per vertical and per region, each with its own dated policy references:

• The country matrix: every category against every major market
• Gambling and sportsbook paid media in the United States
• Gambling advertising across Europe and the UK
• Brazil betting advertising and the rest of LATAM
• Australia's 2027 wagering advertising reforms
• Gambling and casino advertising in India after the 2025 Act
• Dating and companionship advertising across all four platforms
• Crypto and fintech paid media compliance
• Healthcare and telehealth paid media compliance

And if the account is already gone: Meta, Google Ads, TikTok and Snapchat each have a different recovery path, and using the wrong one wastes the window you have.

If you want this run rather than read, our restricted vertical advertising service handles classification, certification and landing page compliance, and ad account recovery covers the diagnosis and appeal when something has already been disabled.

Before any of this is a category problem, it is usually a fundamentals problem. The general guides cover what applies to every advertiser on every platform: why ads get rejected, landing page compliance, claims and copy, advertiser verification and account structure and risk containment.