Copy for Originals · HIPAA conversion tracking · Guide · Healthcare
Yes, as we read Google’s policy: its fertility ad rule bans these ads in a list of locations, the US is not on the list, and its policy page names no certification. The hard limit is targeting. Infertility treatment is one of Google’s examples of its Health category, so remarketing lists, Customer Match and lookalikes are out. Other rules can still apply: prescription drug terms are restricted, and telemedicine that facilitates prescribing needs LegitScript accreditation plus Google certification. Google’s policy also doesn’t settle privacy law: state health data laws and HIPAA have their own sections below.
Every quotation and figure from Google, a statute, HHS, a law firm or an agency on this page was checked word for word against the saved source text on November 2, 2026 (Freshpaint’s line on October 30, 2026). The scan’s counts are our own, from the agent-run scan in the method section. Educational only, not legal advice. No lawyer reviewed this page, and the open questions are listed below. Talk to your own lawyer before acting on it.
As we read it, yes. Google’s “Birth control, fertility, and testing” policy bans these ads in the locations it lists, and the US is not on the list. That is our reading of a list, because the page has no sentence about the US either way. Apart from a general preface, covered below, the page names no certification and no application.
| Rule | What Google’s page says | When it reaches a US clinic |
|---|---|---|
| Fertility policy | No promotion of birth control, fertility and STI products and services in the listed locations. | The US is not on the list. A campaign aimed at a listed location is a different matter. |
| Health in personalized advertising | Advertiser-curated audiences “can't be used if promoting in sensitive interest categories”. “Treatments for pregnancy and infertility” is one of the examples. | Whenever the ad promotes infertility treatment. |
| Prescription drug terms | “Google restricts the use of prescription drug terms in ads, landing pages, and keywords.” | When ads, landing pages or keywords use such terms. The page does not name fertility drugs here. |
| Telemedicine (US row) | Providers “accredited by LegitScript’s Healthcare Merchant Certification Program”, and “Advertisers must also be certified with Google.” | Only if the clinic provides virtual care and facilitates prescribing. |
| Laws and standards | “ads and destinations are expected to follow appropriate laws and industry standards.” | Always. The state and federal sections below say what we could read. |
This guide covers the ad rules. The tracking side, and the tools that send Google less data, are in our HIPAA conversion tracking guide.
Google’s page says it “doesn't allow the promotion of products and services related to birth control, fertility, and sexually transmitted infections (STI) in the locations below.” The locations are Bahrain, China, Djibouti, Egypt, Hong Kong, Iran, Iraq, Jordan, Kuwait, Malaysia, Lebanon, Libya, Morocco, Oman, Palestinian Territory, Qatar, Saudi Arabia, Syria, Thailand, Tunisia, Ukraine, United Arab Emirates and Yemen.
What it covers. Google’s parent healthcare policy gives examples: “condoms, birth control pills, fertility testing, surrogacy services, and STI testing.” IVF is not named, but the examples come after the words “include, but aren't limited to”, so its absence is not an exemption. A clinic that markets to patients abroad should read the list before it adds a location.
Certification. The fertility policy text and the parent policy’s fertility section name no certification or application. Two general lines can make it look otherwise. A preface on Google’s healthcare pages, the fertility page included, says some content can run only “in certain locations by advertisers who have applied and been approved”. And a sentence on the parent page, starting “For most healthcare policies”, says a campaign with a properly certified domain is labeled “Eligible (limited)”. Neither names fertility, so neither tells you this policy needs certification. Google’s fertility page names no certification, and that is as far as we can go.
What happens on a violation. The fertility page says violations “won’t lead to immediate account suspension without prior warning” and that “Google will issue a warning at least 7 days before suspending your account.”
A fertility clinic can’t use Google’s advertiser-curated audiences: Customer Match, your data segments (remarketing lists), audience expansion and lookalike segments. Google’s personalized advertising policy says “Content related to Health is a sensitive interest category”, lists “Treatments for pregnancy and infertility” among its examples, and says these audiences “can't be used if promoting in sensitive interest categories, because they may inadvertently contain sensitive user signals.”
| Targeting feature | Google’s column | For a fertility clinic |
|---|---|---|
| Customer Match | Advertiser-curated | Out |
| Your data segments (remarketing lists) | Advertiser-curated | Out |
| Audience expansion | Advertiser-curated | Out |
| Lookalike segments | Advertiser-curated | Out |
| In-market segments | Predefined | Still available |
| Affinity | Predefined | Still available |
| Life events | Predefined | Still available |
| Demographics and detailed demographics | Predefined, “with exceptions” | Available, with the exceptions Google footnotes |
| Location targeting | Predefined, footnoted | Available; see the geofence point under state laws |
| Custom segments | Predefined, footnoted | Limited, see below |
Why the predefined ones survive. Google says they “can be used even if promoting in sensitive interest categories, because the sensitive user signals are automatically excluded from them.” On that description, none of them should be a list of people Google has marked as seeking fertility care.
Custom segments are the catch. Google says custom segments using sensitive creative or “pointing to sensitive landing pages will only serve with Display campaigns to non-sensitive audiences or contextually”, and that other campaigns using them “won't be eligible to serve.” A clinic’s landing pages are likely to count as sensitive. That is our reading; Google’s page doesn’t mention clinics.
Google’s one-line version. A 2025 Google change log, in a post about content aimed at healthcare professionals, refers to “the existing policy prohibiting targeting users based on sensitive health conditions.”
Google’s table is about audiences. Search keywords aren’t on it, so we draw no conclusion about keywords from it. Remarketing lists is our plain term for what Google’s table calls your data segments; that page doesn’t use the word remarketing.
Take the advertiser-curated audiences out, then build on what Google leaves open to sensitive categories. Google’s first fix for an ad flagged under this policy is: “Review and remove any advertiser-curated audiences from the targeting settings of your ad.” A clinic can do that before any flag arrives.
We don’t say which of these brings more patients. We have no independent data on that, and none of Google’s pages says.
They may, depending on the state and the data, and we read only one such law in full. Washington’s My Health My Data Act lists “Reproductive or sexual health information” as consumer health data, but the word “fertility” appears nowhere in the chapter we read, so its reach to fertility care rests on general wording.
| State | What we read | Reproductive data | Geofence distance | HIPAA carve-out |
|---|---|---|---|---|
| Washington | The statute as published, undated | Named; “fertility” is not | 2,000 feet | Worded for information |
| Connecticut | One 2023 summary by law firm Orrick, with no later check for amendments | Named, per Orrick | 1,750 feet, per Orrick | Worded for entities, per Orrick |
| Nevada | One line in that summary; we could not read the bill | Not read | Not read | Not read |
Washington: what counts. The Act defines reproductive or sexual health information as personal information “relating to seeking or obtaining past, present, or future reproductive or sexual health services”, including “Efforts to research or obtain reproductive or sexual health services”. Those services are “health services or products that support or relate to a consumer's reproductive system or sexual well-being”. Personal information includes “data associated with a persistent unique identifier, such as a cookie ID, an IP address, a device identifier”.
Washington: who. A consumer is “a natural person who is a Washington resident” or one whose consumer health data is collected in Washington. A regulated entity conducts business there, or provides products or services targeted to consumers there, and determines the purpose and means of handling the data. Small businesses are covered on a later date: “A small business must comply with this section beginning June 30, 2024.”
Washington: what it asks. A consumer health data privacy policy; collection only with consent for a specified purpose or as needed to provide what the consumer asked for; sharing only with consent “separate and distinct from the consent obtained to collect”, or as needed for what the consumer asked for; and consent “prior to the collection or sharing”. Disclosure to a processor isn’t “sharing” when it fits the purpose disclosed to the consumer. The statute doesn’t say whether any ad platform is a processor.
Washington: the geofence rule. It is unlawful for any person to put a geofence around an entity providing in-person health care where it is used to identify or track people seeking care, collect their consumer health data, or “send notifications, messages, or advertisements to consumers related to their consumer health data or health care services.” The Act says a geofence “means a virtual boundary that is 2,000 feet or less from the perimeter of the physical location.” Whether a Google Ads radius or location target counts is unsettled: nothing we read mentions Google Ads, and we are not saying a radius is illegal.
Washington: enforcement. The Act says a violation “is an unfair or deceptive act in trade or commerce and an unfair method of competition for the purpose of applying the consumer protection act, chapter 19.86 RCW.”
Connecticut, from one law firm’s summary. Orrick says the amendments define consumer health data to include “reproductive or sexual health data”, and prohibit a geofence “within 1,750 feet of any mental health facility or reproductive or sexual health facility” to identify, track or collect data from a consumer and send them a notification about their consumer health data. Orrick says they took effect on October 1, 2023.
The HIPAA carve-outs differ. Washington’s chapter does not apply to “Information that meets the definition of” protected health information, among other categories: the sentence is about the information, not the organization. Its findings say HIPAA covers “most health care providers” and that data from noncovered entities, “including certain apps and websites”, lacks the same protections. Orrick describes Connecticut’s as “entity-level exemptions”, including for entities governed by HIPAA, so on that summary Connecticut’s rules may not reach a clinic HIPAA governs. We don’t resolve what this means for any one clinic.
Nevada. Orrick calls Connecticut the third state with such protections, after Washington and “Nevada’s new consumer health data privacy law.” We could not read the Nevada bill, so we say nothing about its content.
A federal court vacated most of it on June 18, 2025, according to HHS. HHS’s page says the Northern District of Texas issued an order “declaring unlawful and vacating most of the HIPAA Privacy Rule to Support Reproductive Health Care Privacy”, and that the remaining Notice of Privacy Practices changes “are undisturbed and remain in effect”, with compliance “required by February 16, 2026.”
On what comes next, HHS’s page says: “HHS will determine next steps after a thorough review of the court’s decision.” That is the wording we read on November 2, 2026. The page says nothing about an appeal, and we could not confirm one either way.
The ruling is not permission to track. HHS describes the rule as “prohibiting the disclosure of protected health information related to lawful reproductive health care in certain circumstances.” It was about disclosures of protected health information, and nothing on that page concerns advertising. We read nothing there that tells a clinic what a tag may send, either way.
In an agent-run scan, 11 of 25 sites had a Google Ads tag or ID set up. Of 25 US fertility clinic websites we could read on 2 November 2026, found through OpenStreetMap listings. Not a random sample, not a census: CDC's 2022 list has 423 clinics marked Open.
Agent-run An AI agent ran this scan on November 2, 2026, following the method below. Counts only: no clinic is named, and we publish no per-site data. The totals are in scan-counts.csv.
| What the site’s homepage or Tag Manager file showed | Sites |
|---|---|
| Google Ads tag or ID | 11 of 25 |
| Of those, visible only in the Google Tag Manager container file | 7 of those 11 |
| Google Analytics 4 | 16 of 25 |
| Google Tag Manager | 16 of 25 |
| Meta pixel | 7 of 25 |
| Google Ads tag or ID, or Meta pixel | 11 of 25 |
| A named consent tool | 7 of 25 |
| Freshpaint | 4 of 25 |
One of the 11 is a Google Ads ID held in a Tag Manager variable with no native Ads tag. A further 8 US clinic sites on the map list could not be read: 2 blocked our request, 1 timed out, 1 returned an error, 1 had an unreachable robots file and 3 returned no content. Every count is a lower bound: we read one page per site, without running JavaScript, so tags on booking pages or sent server-side don’t show.
The practical point. 7 of the 11 show up only inside the Tag Manager file, so looking at page source can miss them. To see what your own site has set up, read the container file your page names as well as the page itself.
Freshpaint. It appeared on 4 of the 25. Its pricing page describes its Comply plan with the words “Build a privacy-safe foundation to collect, govern, and activate healthcare marketing data.” A site using it may also send data server-side, which a page scan can’t see.
What the counts don’t show. A tag being set up is all this shows. It doesn’t show that ads are running, what anyone spends or how a tag fires. It says nothing about whether any clinic follows any law, what agreements it has with vendors, or what data reaches Google. A consent tool’s presence or absence says nothing about those either.
An AI agent ran the scan on November 2, 2026, so it is labelled agent-run: it is not a test by a person. It read each site’s homepage, plus up to three public Google Tag Manager container files the page named, with its own user agent, at most one request a second, a 15-second timeout, robots.txt respected and no retry after a block or error. Nothing behind a login, form or CAPTCHA was fetched.
Why a map list. CDC’s 2022 Final ART Services and Profiles list (dataset ix4g-rt8v on data.cdc.gov) has 512 clinics: 423 Open, 54 Non-reporter, 30 Reorganized and 5 Closed. It holds contact details and services but no website column, so it can’t be scanned directly. SART’s clinic finder is a location search with no list file, and reading its back end would be scraping behind a form. So the starting list came from OpenStreetMap, which is volunteer-mapped open data (ODbL), through the Overpass API: places whose speciality tag names fertility, IVF or reproductive endocrinology. That returned 212 elements, 150 with a website of their own, and 78 distinct hosts once social, directory and booking hosts were dropped and places sharing a website merged. Two further queries by place name failed on the public server and were dropped, so clinics mapped without the speciality tag are missing.
| Step | Removed | Left |
|---|---|---|
| Hosts on the map list | 78 | |
| Outside the US (the query’s box also covers Canada and Mexico) | 17 | 61 |
| Not fertility clinics (tagged on the map, but other kinds of business or a wrong website: acupuncture, egg donor banks, a psychologist, a skincare shop, a supplement retailer, a domain now serving a casino page and others) | 26 | 35 |
| Chain duplicates (3 hosts that redirect to one group, counted once) | 2 | 33 |
| Could not be read (2 blocked, 1 timed out, 1 error, 1 unreachable robots file, 3 stubs or a bot challenge with no content) | 8 | 25 |
What counted as each tag. A first pass miscounted Meta pixels, so the sites that returned a page were read again the same day with the corrected matching below, and the counts come from that read.
AW- ID in the page or container, or a native Ads conversion, remarketing or related tag type in the container (a conversion linker alone isn’t counted). Paused tags aren’t counted.G- ID in the page or container, or a GA4 tag type in the container. Google Tag Manager: a GTM- ID in the page.What we left out. The scan also looked for a few other data and privacy tools by name. We report only Freshpaint. A missing name proves nothing, because such tools can sit on a clinic’s own domain, and a named tool found on fewer than 3 sites isn’t shown. We publish whole-sample totals and no site list, so the figures alone don’t single out a clinic.
What it can’t see. Tags that load only after a consent click or by script the HTML doesn’t name, tags on pages other than the homepage, server-side tagging, and anything behind a login. Sites that blocked the request may not be a random set.
Why there are no per-site results here. Our disclosure page says agent-run checks come with the date, the inputs and the raw outputs unless that would identify a third party, and our other agent-run test publishes them. This scan gives the date, the rule that picked its inputs and its totals in scan-counts.csv, but not the list of sites or the per-site results. Those would identify clinics: the map list is public, so many rows could be matched to a clinic, even without names, by scanning the same sites again. So we hold those back, and you can’t check the totals against our site-level results.
A random draw from CDC’s Open clinics, with websites looked up one by one, would remove the map list’s bias. This scan isn’t that, and it can’t speak for US clinics as a whole.
We found no independent figure. Everything we found is one agency’s page relaying others, with no sample stated, and that agency sells this service.
| Credited to | Monthly figure, as Web Tonic gives it | What kind of figure |
|---|---|---|
| Tandem Medical Marketing | “$10,000–$20,000 monthly” in major metros; “$5,000–$8,000 per month” in smaller metros | Said to be typical in major metros and viable in smaller ones; no spend data shown |
| Medical Marketing Firm | “$2,000–$6,000 per month for Google Ads” within a “$8,000–$20,000 total marketing budget” | A recommendation, not measured spend |
The same page also gives an average click and a cost per lead. We don’t repeat them: neither comes with a sample, and its cost per lead differs from one of its tables to the next. Use your own account’s numbers once it has any.
Three questions stay open, and this guide doesn’t answer them: whether a Google Ads radius or location target counts as a “geofence”; whether a clinic’s ad-click data is protected health information (exempt in Washington) or consumer health data (covered); and whether Google is a “processor” under Washington’s Act. Nothing we read applies these rules to a clinic’s ads. They are the ones to take to your own counsel.
We manage Google Ads and do not give legal opinions. If you want campaigns built inside the limits above, see Copy for Originals’ services.
Each source was read on November 2, 2026, except where a different date is given. Dates are the sources’ own; most of these pages show none.