In January 2022, the Center for Intimacy Justice published a study of 60 organisations working in women's health. Every single one of them had had an advertisement rejected by Meta. Around half had had an entire advertising account suspended at some point. Fifty-nine of the 60 were founded and led by women.
The rejected campaigns were not fringe. They covered menopause, pelvic pain, pregnancy, postpartum recovery, menstrual health, fertility and sexual wellness. Ads were classified as adult content for using the word vaginal. Meanwhile, advertisements for erectile dysfunction ran without difficulty.
Nine months later, Meta revised its policy. The updated guidelines specifically named products addressing the effects of menopause, pain relief during sex, and sex education as permitted. The examples Meta added were, almost exactly, the examples the report had shown as blocked.
Then, in its 2025 report, the same organisation found that 84 percent of the businesses it surveyed had still had ads rejected on Meta.
That is the part worth sitting with. The policy changed. The outcome did not. And in the six years between the first complaints and now, something happened that almost nobody planned: an entire category quietly rebuilt the way it acquires customers, one workaround at a time, without ever deciding to.
What actually happened, in order
Women's health advertising restrictions are not a single event or a single platform decision. They are a six-year sequence in which each attempted remedy was followed by evidence that the underlying behaviour continued.
| When | What happened |
|---|---|
| 2020 | Femtech founders begin going public about rejections. In reporting from that year, one brand described having to be about describing its own products to avoid triggering review. |
| January 2022 | The Center for Intimacy Justice report lands, produced with the pelvic floor physical therapy company Origin. Sixty organisations, 100 percent experiencing rejection, roughly half suspended. |
| October 2022 | Meta revises its sexual health advertising policy and adds permitted examples. The organisation's founder, Jackie Rotman, notes publicly at the time that the open question is whether enforcement will follow the policy, or whether the algorithms will keep rejecting. |
| July 2023 | The Center for Intimacy Justice files a complaint with the Federal Trade Commission, alleging that Meta claims to permit these advertisements while systematically rejecting them. Five members of Congress publish a letter asking the FTC to review it. The complaint documents rejections of advertisements using the exact language Meta's own revised policy lists as allowed. |
| January and February 2025 | A different kind of restriction arrives, and this one has nothing to do with vocabulary. Meta introduces its sensitive ad categories. Businesses categorised as health and wellness lose the ability to optimise campaigns for lower-funnel conversion events including purchases, add-to-cart, checkout completions and booked appointments. Custom audiences built on sensitive health behaviours are disabled. Lookalike audiences derived from them are prohibited. Advertisers are directed toward upper-funnel objectives such as landing page views and video views instead. Misclassification is common, and appeals take several days to process. |
| 2025 | The second Center for Intimacy Justice report widens the lens to four platforms and 159 organisations serving people in more than 180 countries. Eighty-four percent report ads rejected on Meta. Sixty-six percent report ads rejected on Google. Sixty-four percent report product listings removed on Amazon. Respondents estimate annual revenue losses running from 10,000 dollars to 1 million dollars per company on Amazon alone. In a parallel survey by CensHERship, 95 of 95 respondents reported at least one censorship issue, and 17 percent reported as many as ten. |
| 2026 | A second wave of Meta restrictions extends into healthcare lead generation. Form fields that could carry protected health information get flagged, blocking lead and appointment-scheduling events. Direct links to patient portals are restricted. The campaigns clinics and telehealth providers use to book appointments are now inside the scope. |
Six years. Two reports, one FTC complaint, one Senate letter, one policy revision, and two rounds of structural tightening that had nothing to do with the original complaint at all.
Why are women's health ads rejected? There are two separate answers, and they get conflated
Women's health brands face two distinct advertising problems, and treating them as one problem is why so much of the response to them has been ineffective.
The first problem is rejection. You cannot reliably say the words. Clinically accurate language for anatomy, symptoms and conditions triggers adult-content classification. This problem is well documented, widely reported, and has produced a decade of creative gymnastics: euphemism, abstract imagery, general-wellness framing in place of the specific claim that would actually persuade someone.
The second problem is measurement, and it is newer. Even when the advertisement runs cleanly, a brand classified as health and wellness has lost the machinery that made paid social work in the first place. No lower-funnel optimisation. No conversion-based audiences. No lookalikes of purchase behaviour. The campaign is still ongoing. It just spends against upper-funnel proxies while the signal that told you whether it worked has been switched off.
The second problem hits a brand the first one never touched. A funded femtech company with a careful compliance function, a clean ad account and a six-figure monthly budget may have gone years without a single rejection. In 2025 that company was reclassified, lost lower-funnel optimisation, and watched efficiency drift with no policy violation to appeal and no rejection notice to point at.
This is the version of the story that does not get told, because it is not an injustice narrative. It is a measurement narrative. And it is the one currently costing funded brands the most money.
What six years of this actually built
The lasting consequence of advertising restrictions in women's health is not the money lost to rejected campaigns. It is that the category developed a community-dependent go-to-market model by accident, and therefore built no infrastructure for it.
Watch what brands did as each door closed. They rewrote copy into euphemism. They routed spend to influencers, which moved the claim from an ad account into a creator's mouth without moving the liability, since under the endorsement rules the brand still carries primary responsibility for what is said on its behalf. They leaned harder on organic and email. They hired community managers. They showed up in the subreddits and the condition-specific groups where their customers were already comparing notes.
Each of those was a sensible tactical response to a specific blocked door. Collectively they amount to a channel strategy that nobody wrote down.
Which is why femtech now sits in an unusual position. It is a category where the purchase is genuinely researched, where trust decides the outcome, and where a large share of that research happens in public community threads. Estimates of the category's size vary by an order of magnitude depending on the source, from under 10 billion dollars to over 70 billion dollars for 2026, which tells you the analysts are not sure either. But the direction of the money is not in dispute, and it is flowing into a category whose main acquisition channel was never designed, never resourced properly, and is usually still owned by whoever had spare capacity.
The part nobody planned for
The communities that women's health brands were pushed into by ad restrictions have since become the source material for AI-generated answers, which means the improvised channel turned into the discovery channel.
Independent and vendor analyses through 2026 agree on the direction even where they disagree on the magnitude: community discussion now supplies a substantial and growing share of the citations that AI answer engines use, with Reddit consistently among the most-cited domains. The share varies sharply by engine, and it moves month to month, so anyone quoting a single confident percentage is quoting an average of unstable things. But the structural point holds. When a woman asks an AI assistant which fertility clinic protocol to ask about, or whether her symptoms sound like perimenopause, a meaningful part of the answer is assembled from community threads.
Combine that with what people are actually doing. OpenAI reported in January 2026 that more than 40 million people ask ChatGPT healthcare questions daily, and that most health conversations happen outside standard clinic hours.
So the sequence completes itself. A category is pushed out of paid channels. It improvises its way into the community. Community becomes the layer that AI answers are built from. The improvised channel is now the front door, and it is still being run as an overflow valve.
What follows from this
Four things follow, and none of them are creative fixes.
- Treat community as a channel, not as a workaround. A channel has an owner, a budget, a cadence, a quality standard and a reporting line. Most femtech community work has volunteers and good intentions. That gap is the whole opportunity.
- Measure it with an instrument that can read it. Last-click attribution structurally cannot see a woman who reads a thread in March, remembers the brand in June, and arrives as direct traffic. Practitioner research on dark-social channels finds a gap of roughly 90 percent between what software attributes and what customers report when asked directly. One question on the signup form recovers more of this than any tracking configuration will.
- Stop fighting the vocabulary gap and use it. The words that survive ad review are not the words your customer types. Community is the one channel where a brand can use the same language as the person it is talking to. That is not a compliance risk to be managed. It is the reason the channel works.
- Put a named human on every sentence. This is where the endorsement rules, the platform rules and simple decency converge. Reddit spent 2026 tightening enforcement against automated and inauthentic participation. The FTC extended its endorsement guidance to cover AI-generated content, and the brand carries primary liability regardless of what drafted the words. In a category where the audience is often frightened and frequently dismissed by clinicians, one tone-deaf reply does damage that no impression volume repairs. A person who knows the brand, reading every response before it posts, is not cautious. It is the minimum defensible way to operate here.
A note on what we do
We build and run the Ayana Agent for brands in exactly this position. It monitors the communities where a category's decisions actually get made, drafts a reply in the brand's voice with genuine command of the product detail, and routes every draft to a human on the client's team who approves, edits or discards it before anything is posted. Nothing goes out unread. We started in maternal health, under one of the stricter marketing codes in existence, because that is where getting it wrong costs the most.
The restrictions are not going to lift. The chronology above is six years of evidence for that. The question is whether the channel they created gets run deliberately or accidentally.
See whether your category is one we can help with.
A fit call is a short, direct conversation about your communities, your compliance constraints, and whether a human-approved agent is the right instrument for the channel.
Frequently asked questions
This article describes advertising policy and regulatory context. It is not legal advice.