Meta and Google closed the door on cause-driven advertising.

A homelessness charity writes an ad. The copy says a few thousand people slept rough in this city last winter, please donate. The ad goes to a donation page. Meta rejects it. The reason given is that homelessness is a social issue.

That is not a moderation error. It is the policy, and it has been the policy since 6 October 2025, the day Meta stopped selling political, electoral and social issue ads anywhere in the European Union. Google got there first, announcing its own EU political advertising withdrawal on 14 November 2024.

Almost everything written about this was published before it happened, in the future tense. This is the retrospective. What actually changed, what the scope really is, who it hit, and what still works. The short version: every rule that broke advertising on Meta regulates the use of personal data to select an audience. Contextual advertising selects a page. Those rules have nothing to bite on.

What exactly did Meta and Google change, and when?

Two separate restrictions, nearly four years apart. On 19 January 2022, Meta removed detailed ad targeting options covering sensitive topics, so cause-driven advertisers could no longer find their audience. On 6 October 2025, Meta stopped selling political, electoral and social issue ads in the EU altogether, so they could no longer run the ad at all. Google announced its own EU withdrawal on 14 November 2024.

Wave one was announced in late 2021 and took effect on 19 January 2022, with existing campaigns allowed to keep delivering until 17 March 2022. Meta removed detailed targeting options relating to health causes (its own examples included "Lung cancer awareness" and "World Diabetes Day"), sexual orientation ("same-sex marriage", "LGBT culture"), religious practices and groups ("Catholic Church", "Jewish holidays"), race and ethnicity, and political beliefs, social issues, causes, organisations and figures. Meta's stated reason, per Search Engine Land's report, was that these options "relate to topics people may perceive as sensitive". The change applied across Facebook, Instagram, Messenger and Audience Network. Meta cited feedback from civil rights experts and policymakers, and published separate guidance for nonprofits.

Wave two is the one that matters now. The EU's Transparency and Targeting of Political Advertising regulation, or TTPA, is Regulation (EU) 2024/900: a law setting transparency, labelling and targeting rules for political advertising across the Union, with penalties reaching 6% of annual turnover. It entered into force on 9 April 2024 and applied from 10 October 2025. Meta announced on 25 July 2025 that it would stop allowing political, electoral and social issue ads in the EU from 6 October 2025, across Facebook, Instagram and Threads. In its own announcement, Meta said the regulation left it with "an impossible choice: alter our services to offer an advertising product which doesn't work for advertisers or users, without guarantee that our solution would be viewed as compliant, or stop allowing political, electoral and social issue ads in the EU".

Google did not follow Meta. Google went first. This is the detail most coverage gets backwards. Google's announcement is dated 14 November 2024, eight months before Meta's, and covers Google Ads and YouTube in the EU. Google wrote that the TTPA "introduces significant new operational challenges and legal uncertainties", and that it "defines political advertising so broadly that it could cover ads related to an extremely wide range of issues that would be difficult to reliably identify at scale". Google also deleted its EU political ads transparency archive, removing roughly seven years of records across all 27 member states, a change spotted by 404 Media on 28 September 2025. The European Commission adopted the implementing act for its own political ads repository on 9 April 2026, which left a transparency gap of roughly six months.

What counts as a "social issue" ad on Meta?

In the EU, Meta applies eight social issue categories, set out in its advertising standards for social issues, elections and politics: civil and social rights, crime, economy, environmental politics, health, immigration, political values and governance, and security and foreign policy. Outside the EU the same list adds education and guns, making ten. An ad does not have to name a party, a candidate or an election. Touching one of these topics is enough.

This is the part that catches people out, so here is the reference table in one place.

The classification follows the topic, not the organisation. According to fundraising platform iRaiser's analysis, the activities routinely caught include driving traffic to an online donation page, promoting a crowdfunder, recruiting participants for a fundraising event, and boosting a peer-to-peer campaign. iRaiser, citing a European Fundraising Association survey, puts the exposure at over 60% of European nonprofits investing in social media advertising, with Facebook and Instagram the most popular platforms. The Charities Institute Ireland reached the same conclusion for its members.

Who did the EU political ad ban actually lock out?

Five groups, none of them political parties. Charities and NGOs running fundraising and awareness campaigns. Advocacy organisations. Independent and investigative media promoting their reporting. Values-led brands whose entire market positioning sits inside one of the eight categories. Health and harm reduction organisations, which are caught by the health category regardless of what they are actually saying.

The asymmetry is the point. A commercial advertiser selling furniture, software or holidays was barely affected by any of this. Their ads do not touch civil rights, immigration or environmental politics, so wave one cost them a few targeting options and wave two cost them nothing at all. The organisations that lost the most access were, almost without exception, the ones with the smallest budgets and the least ability to replace a channel. A well-funded brand can buy its way around a restriction. A charity with a four-person communications team cannot.

And the restriction hits by delivery, not by domicile. A UK or US charity that wants to reach supporters in Ireland, Germany or the Netherlands is inside the ban even though it sits outside the EU.

Was this Meta's fault or the EU's?

Reasonable people disagree, and it is worth being fair here. Meta did not ban these ads out of hostility to charities. It withdrew rather than build compliance machinery for the TTPA, and it said so plainly: an "impossible choice" between an ad product that "doesn't work for advertisers or users" and stopping the category. Google's objection was more specific and, frankly, hard to dismiss: the TTPA "defines political advertising so broadly" that reliably identifying covered ads "at scale" was the problem. The regulation was written to stop opaque political microtargeting and foreign interference. Charity fundraising was collateral, not the target.

None of which changes anything for the organisation that has been locked out. Whether the cause was an overbroad regulation or a platform unwilling to absorb the compliance cost, the effect on the homelessness charity is identical: the ad does not run. This article is about the effect.

Does the ban apply in the UK, Switzerland or the wider EEA?

No to the UK and no to Switzerland. Meta's 25 July 2025 announcement says the restriction covers the European Union and states that "our decision is specific to the EU". The UK is not an EU member state and is not covered. Switzerland is in neither the EU nor the EEA and is not covered. Most competing articles are loose about this, so be careful which one you are reading.

Two caveats worth stating precisely. First, some sector commentary describes the restriction as EEA-wide, which would pull in Norway, Iceland and Liechtenstein. Meta's own wording says EU. If you operate in an EEA state that is not an EU member, verify your position directly rather than trusting a summary, including this one. Second, a UK-based charity is only outside the ban while it is targeting UK audiences. Point the same campaign at EU users and the ban applies, because what is restricted is where the ad is delivered.

This is regulatory territory and it moves. Nothing here is legal advice, and any organisation with real exposure should check its own position with someone qualified to give it.

Why doesn't the TTPA affect contextual advertising?

Because there is nothing for it to regulate. The TTPA restricts the use of personal data to target and deliver political advertising. Contextual advertising selects the page rather than the person: the ad is chosen to match the article's subject matter, with no user profile, no cross-site identifier and no special category data involved. Remove the profile and the targeting rules have no subject matter to attach to.

The same logic runs through every restriction in this story. Wave one removed targeting options that inferred a person's health, religion, ethnicity, sexuality or politics. Contextual never inferred any of them. GDPR consent obligations attach to processing personal data. Contextual does not process any. The rules that broke platform targeting are all rules about profiling people, and the whole point of contextual is that it does not.

Then there is the part cause-driven advertisers tend to miss. The ad industry treats contextual as the degraded option, the thing you fall back on when a cookie is not available. For this kind of advertising that gets the ranking exactly backwards. On Meta you were guessing: you picked a behavioural proxy and hoped it correlated with people who care about climate, or refugees, or press freedom. Contextually, you place the ad beside an article on that exact subject, read by someone who chose to open it. Intent beats inference. Somebody reading about Gaza right now is a better prospect for a Palestinian soap co-op than any behavioural segment ever assembled, and no profile was needed to find them.

The machinery is simpler too. Real-time bidding, or RTB, is the auction system where a single ad impression is offered to many buyers in milliseconds, passing through exchanges, demand-side platforms, supply-side platforms and verification vendors before it clears. Each hop takes a cut. Fixed-price contextual is one call, one decision, one ad. That difference shows up in three numbers worth knowing:

• Consumer Reports examined the Facebook data archives of 709 volunteers in January 2024 and found 186,892 companies had sent data about them to the platform, an average of 2,230 companies per person.
• Scope3 measured programmatic advertising at more than 215,000 metric tonnes of carbon emissions per month across the US, UK, France, Germany and Australia, and found that cutting the worst-emitting 10% of domains would remove 33,500 tonnes of that every month.
• The ISBA and PwC Programmatic Supply Chain Transparency Study, published on 6 May 2020, found that publishers received 51% of advertiser spend on average, and that 15% of spend was an "unknown delta" that could not be attributed to anyone at all. The second study, published 18 January 2023, cut that unattributable share to 3%. Real progress, and still roughly half of every budget consumed before it reaches the person who made the thing you advertised next to.

Why do brand-safety blocklists compound the problem?

Because they cut the other side of the same match. Brand safety blocklists are keyword lists advertisers use to keep ads away from content they consider risky, and they are blunt. Words like refugee, conflict, abuse, protest and death are common entries. Those words are not a description of unsafe content. They are a description of journalism.

So the publisher covering an issue seriously gets demonetised for covering it, and the advertiser who exists to campaign on that issue gets blocked for caring about it. Both halves of a genuinely good match are severed by the same machinery, and neither party did anything wrong. A newsroom reporting on famine loses revenue on its best work. The aid organisation that wants to reach exactly those readers is told the placement is unsafe.

The mechanism is documented by Google itself. Advertisers can exclude specific webpages, domains and topics from their campaigns. On the publisher side, Google Publisher Restrictions state plainly that when content carries an inventory restriction, fewer advertising sources are eligible to bid on it, and in some cases no ads appear at all. The publisher is not told which advertisers excluded them, so there is no notice and therefore nothing to appeal. They simply see revenue that is not there.

The effect on the ground is that serious publications stop trying. Dave Reed at Mondoweiss, one of our publishers, put it to us unprompted: "We have experienced platforms restricting us in the past, and I expect most of the major ad networks would do the same, so I haven't tried to set them up." That is what a decade of silent exclusion produces. Not an appeal, just a publication that assumes the answer and never asks the question.

That gap is what Ethicly was built in, and it is why we run brand safety as a two-way allow list rather than a keyword blocklist. Publishers choose which advertiser categories they will host. Advertisers choose the publications by name. Both decide in advance and a person reviews both sides. It does not scale the way an automated blocklist scales. It also does not defund a newsroom for using the word "protest". The longer version of that argument, with sources, is on our about page.

What are the alternatives to Meta ads for charities in the EU?

Six that work, in rough order of how quickly you can stand them up. None of them individually replaces the reach Meta had, and anyone telling you otherwise is selling something.

• Organic on Meta. Start here, because it is free and it still works. Meta's announcement is explicit that the ban "won't prevent people in the EU from continuing to debate politics on our services". You lose paid amplification, not the account.
• Contextual ad networks. Ads matched to page content on publishers whose audiences already care. No profiling, so no TTPA targeting exposure.
• Direct deals with publishers. Slower to negotiate, but you get guaranteed placement, a real relationship, and a rate card you can plan against.
• Newsletter sponsorship. Often the best value in this whole list for cause-driven advertisers. Engaged, self-selected audiences, and a sponsor slot in a newsletter about your issue is about as contextual as advertising gets.
• Podcast and audio. Host-read sponsorship carries trust that display cannot buy, and audio has largely avoided the brand safety blocklist problem.
• Owned channels. Email and SEO. Unglamorous, slow, and the only channels no platform can take away from you on ten weeks' notice, which is all the warning Meta gave between its announcement and the cutoff. The past four years should have made the case for owning your distribution.

Where does Ethicly fit?

Honestly, and at our current size. Ethicly is a fixed-price contextual ad network based in Switzerland, built specifically for the publishers and advertisers the large platforms exclude. As of August 2026 we have four live publisher sites: Mondoweiss, Bulletin, Find A Protest, and our own Tabs for Palestine extension. More than two dozen publications sit in the pipeline behind them. Five advertisers have come back for a second campaign. That is a small network, and you should hear it from us rather than find out later.

What we can do today: run contextual campaigns on independent publications, at $2 to $5 fixed CPM, no minimum spend, creative included. No auction, so you know the rate before the impression instead of after. You choose the publications by name, and the publisher approves your creative before it runs. Publishers keep 80% of gross. Each publisher sets its own rate, and our pricing page has a calculator to size a campaign against it.

One thing to be straight about, because it would be easy to read this article as a pitch for political advertising. It is not. Political campaigning is the first entry on the list of categories we refuse at network level, alongside gambling, tobacco, alcohol, cryptocurrency, and military and defence. We do not run ads for parties, candidates or electoral campaigns, and no budget changes that.

The difference is that we do not treat a homelessness charity as a political campaign. Meta's eight EU categories collapse those two things into one classification. A donation page for winter shelter becomes "economy". A disease awareness day becomes "health". That collapse is the error this whole article is about. Refusing party political advertising and refusing to advertise a food bank are not the same decision, and a network should not need the same rule for both.

What we cannot do: match the reach of a platform with billions of users, or promise you performance we have not measured. We are one of several contextual networks and you should look at the others too. An organisation that has just been shut out of two platforms does not need a third one overpromising. If you are a publisher rather than an advertiser, the same argument from your side is on our publisher page.

The ad still needs to run

The homelessness charity from the opening still has a donation page, still has a winter appeal, and still needs people to see it. What changed on 6 October 2025 is not whether that ad is worth running. It is which door it goes through. Contextual advertising never needed Meta's permission, because it never needed anyone's profile.

If you have had a campaign rejected as a social issue ad, tell us what you were trying to run and we will tell you honestly whether we can help. Plan a campaign, or read more about why contextual works without profiling.