Summary

  • We scanned 94 Australian clinic websites in October 2026. 63% fired at least one advertising tag before the visitor had any chance to consent.
  • 18% sent a page address naming a condition or treatment to an ad platform, and another 18% sent a booking-page address.
  • Meta now strips page addresses on almost every health site we saw. The exposure has moved to Google Ads, Microsoft Bing, Pinterest, Snap, LinkedIn and Reddit tags.
  • Only 19 of 94 sites showed a cookie banner, and 12 of those 19 still fired ad tags first.
  • This is a pilot: a hand-built sample, one visit per site. It describes what a browser saw, not whether any clinic broke the law.

In June 2026 the Australian Privacy Commissioner found that two health providers interfered with the privacy of website visitors by using third-party tracking pixels without consent1. Alongside those decisions, the OAIC published the results of its own inspection of 50 health service provider websites, which found that 52% used a third-party tracking pixel2.

We wanted to know what an ordinary patient’s browser encounters on Australian clinic websites today, four months later. This is the first edition of the Akesis Clinic Pixel Index: a pilot scan of 94 sites across six types of practice. No clinic is named in this report, and none will be.

Key findings

63%fired an ad tag on page load, before any consent interaction (59 of 94)
18%sent a condition- or treatment-named page address to an ad platform (17 of 94)
20%showed any cookie banner or consent tool (19 of 94)

1. Most clinic sites load advertising tags straight away

On 59 of the 94 sites, at least one advertising tag sent a request to an ad platform as soon as the page loaded. Our scanner never clicked a banner, form or button, so every one of those requests happened before the visitor made any choice. Google Ads or DoubleClick tags fired on 48% of sites and the Meta pixel on 44%. Google Analytics 4 was present on 77% and Google Tag Manager or the Google tag on 82%.

Session replay tools, which record how a visitor scrolls, clicks and types, ran on 28% of sites.

2. The leak has shifted from Meta to other platforms

Meta told advertisers that data sources in sensitive categories can be restricted, and that its most common restriction, Core Setup, “restricts the sharing of custom parameters and parts of URLs following the domain”3. Our data is consistent with that. On 38 of the 40 sites where the Meta pixel fired on an inner page, the page address it received contained only the domain. Meta received a full inner-page address on just 2 of 94 sites.

Other ad platforms were not stripping addresses in the same way. On 17 sites, a Google Ads, Microsoft Bing, Pinterest, Snap, LinkedIn or Reddit tag received an inner-page address that named a condition or treatment. On another 17, the address was a step in the booking process. Counting generic service pages too, 41% of sites sent an ad platform an address from a service, condition or booking page.

Why this matters for your own checks: a review that only looks at the Meta pixel would have cleared most of these sites. Meta also says its systems “are not a substitute for your own compliance mechanisms”3.

3. Cookie banners rarely stop the pixels

Only 19 sites showed any banner or consent tool. Cookiebot was the most common (9 sites), followed by Usercentrics and OneTrust (2 each). On 12 of those 19 sites, advertising tags fired before the visitor had clicked anything. The other 75 sites had no banner at all. The OAIC has said that steps to notify visitors could include a banner or pop-up at the first visit2, so a banner that does not control what loads is a gap worth closing.

4. Groups carry more tags than independents

Multi-site groups were far more heavily instrumented: 80% fired an ad tag before consent (independents: 48%), 70% fired the Meta pixel (20%) and 48% ran session replay (10%). Groups also sent condition or treatment addresses more often (23% against 14%). A scan cannot tell us why, but it is consistent with groups running more paid campaigns across more platforms, where each new campaign can add another tag.

5. Privacy policies often don’t mention ad tracking

Our automated check could not find a privacy policy on 21 sites (22%). Of the 59 sites with an advertising tag, 25 had a policy that never mentioned advertising, remarketing, pixels, Facebook or Meta. Of the sites loading the Meta pixel, 55% never mentioned Meta, Facebook or pixels. This closely echoes the OAIC’s own inspection, which found 77% of the pixel-using sites it checked did not mention pixels in their privacy policy2.

Results by type of clinic

Ad tags before consent, and condition or treatment URLs sent to ad platforms

Ad tag fires before consentCondition or treatment URL sent to an ad platform

Ad tags firing before consent, and condition or treatment URLs sent to ad platforms, by type of clinic Cosmetic/skin (14): 86% fire an ad tag before consent, 29% send a condition or treatment URL; IVF/fertility/specialist (9): 78% fire an ad tag before consent, 44% send a condition or treatment URL; Physio/allied (17): 65% fire an ad tag before consent, 6% send a condition or treatment URL; Dental (18): 61% fire an ad tag before consent, 17% send a condition or treatment URL; GP/medical (24): 54% fire an ad tag before consent, 8% send a condition or treatment URL; Psychology/mental health (12): 42% fire an ad tag before consent, 25% send a condition or treatment URL; All clinics (94): 63% fire an ad tag before consent, 18% send a condition or treatment URL. 0% 25% 50% 75% 100% Cosmetic/skin (14) 86% 29% IVF/fertility/specialist (9) 78% 44% Physio/allied (17) 65% 6% Dental (18) 61% 17% GP/medical (24) 54% 8% Psychology/mental health (12) 42% 25% All clinics (94) 63% 18%
Per cent of sites in each segment. Number of sites in brackets. Segments of 9 to 24 sites are small, so treat differences between them as indicative only. Source: Akesis Clinic Pixel Index 2026 pilot.

Cosmetic and skin clinics were the most heavily tagged: 86% fired an ad tag before consent, 79% loaded the Meta pixel and 64% ran session replay. Six of the eight sites with a TikTok pixel were cosmetic chains. Fertility and specialist sites had the highest rate of condition or treatment addresses reaching an ad platform (44%), although this segment had only nine sites. Psychology and mental health sites had the fewest ad tags overall (42%), but a quarter still sent a condition-named or booking address to an ad platform. That matters because the OAIC has suggested that where simply visiting a website reveals sensitive information, such as a mental health service, it may be preferable not to use pixels at all4.

Figures are % of sites in each segment. N = number of sites.

Pixel Index results by segment, per cent of sites
SegmentNAd tagMetaGoogle AdsPre-consentBannerReplayCondition URLBooking URLSilent policy
All clinics94634448632028181827
GP / medical245429295417298425
Dental18613361612822171122
Physio / allied1765475365122462941
Psychology1242254242170252533
IVF / fertility978674478442244220
Cosmetic / skin14867964861464292929
Multi-site groups44807059802548232527
Independents50482038481610141226

All figures are per cent of sites in the segment. “Pre-consent” means an ad tag fired before any consent choice; “Replay” means session replay. Psychology includes mental health services, and IVF / fertility includes other specialist clinics. “Silent policy” means the site fired an ad tag and its privacy policy, where found, did not mention advertising tracking.

What one site looked like

To show what these numbers mean in practice, here is one anonymised example. An independent mental-health practice sent the address of a page naming a specific condition to Google Ads as soon as the page loaded. The same site also loaded Meta and LinkedIn tags, showed no consent banner, and had a privacy policy that did not mention advertising tracking. There is no suggestion that anyone at the practice intended this, and a scan cannot show how a setup came about. The OAIC has described the same pattern more broadly: many organisations it engaged with were not aware of all the tracking pixels on their websites, often because marketing sat apart from privacy or was outsourced25.

Booking platforms

We detected an online booking platform on 41 sites: HotDoc (13), AutoMed (7), Centaur/D4W (6), Cliniko (6), Timely (4), HealthEngine (4), Zanda (3), Nookal (1) and Halaxy (1). Of those 41 sites, 23 also fired advertising tags. We did not scan inside booking widgets hosted on the platforms’ own domains, so these figures cover the clinic’s website only. Our booking platforms guide explains where the risk sits.

Methodology

  • Sample. 106 clinic websites were selected by hand: 26 GP and medical, 18 dental, 18 physiotherapy and allied health, 15 psychology and mental health, 11 IVF, fertility and specialist, and 18 cosmetic and skin. 47 belonged to multi-site groups and 59 were independent. All states and territories were represented. Sites came from web searches for local clinics in named suburbs, plus well-known multi-site chains.
  • Scan. 94 sites scanned successfully. The other 12 refused connections from our scan network or returned errors. Each site was loaded once in headless Chromium with a normal desktop browser identity and Australian English locale. We loaded the homepage, then up to three inner pages linked from it (a service or condition page, a booking page and a contact page). We waited for network activity to settle plus five seconds, and logged every request.
  • No interaction. The scanner never clicked a cookie banner, form or booking button. Any advertising request we logged was therefore sent before the visitor had made any consent choice.
  • Definitions. “Ad tag” means a request to Meta (facebook.com/tr), Google Ads or Floodlight conversion and remarketing endpoints, TikTok, LinkedIn Insight, Microsoft UET, Pinterest, Snap, X or Reddit. Google Ads requests were not counted where Google Consent Mode signalled that ad storage was denied. “Condition or treatment URL” means an ad request whose payload contained the address of an inner page whose path names a condition or treatment.
  • Privacy policies were found from homepage links or common paths, then checked for keywords.

Limitations

  • This is a pilot. The sample was built by hand and is not random. Segments rest on 9 to 24 sites, so small differences between them should not be over-read.
  • Each site was visited once, from a US-based cloud server, using a headless browser. Banners that only show to Australian visitors, bot protection, A/B tests or tags that load later could all behave differently for a real patient.
  • Server-side tracking, such as Meta’s Conversions API or a server-side Tag Manager container, cannot be seen from the browser. Our figures are therefore more likely to understate exposure than overstate it.
  • Inner pages were chosen automatically. 85 of 94 sites had a service or booking page scanned. Booking widgets hosted on third-party domains were not scanned inside.
  • “Fires before consent” is a technical observation, not a legal finding. Whether a particular setup breaches the Privacy Act depends on context, including the notices given and any consent obtained elsewhere.
  • Keyword checks on privacy policies can miss unusual wording. “Not found” means not found from homepage links or common paths.

What to do with these findings

Health service providers are covered by the Privacy Act regardless of turnover6, so the small business exemption will not usually help a clinic. The OAIC’s guidance is clear that the Privacy Act does not ban pixels, but that organisations must understand what they collect, and must have consent before collecting sensitive information such as health information4. A sensible first pass takes less than an hour:

  • List every tag on your site, including ones added by past agencies. Check Tag Manager, your website plugins and any code pasted into page headers.
  • Check what each ad tag receives on a condition, treatment or booking page, not only on the homepage. Include Google Ads, Microsoft, Pinterest, LinkedIn, TikTok, Snap and Reddit, not just Meta.
  • If you use a banner, confirm that advertising tags stay silent until the visitor agrees.
  • Read your privacy policy as a patient would, and check that it names the advertising tools you actually use.
  • Re-check after every campaign launch or website change. The OAIC warns that pixels are not “set and forget”2.

Media and research enquiries about the Index methodology are welcome via our contact page. We will not identify individual clinics.

See what your own website shares

The free Akesis scan checks your homepage for advertising and analytics tags, whether they fire before consent, and what they receive. It takes about a minute and needs no install.

Run a free scan

Sources

  1. OAIC, Privacy Commissioner finds privacy breaches in third-party tracking pixel investigation (media release), 24 June 2026. www.oaic.gov.au/news/media-centre/privacy-commissioner-finds-privacy-breaches-in-third-party-tracking-pixel-investigation
  2. OAIC, Your life, pixelated: how tracking pixels watch your every click, 24 June 2026. www.oaic.gov.au/news/blog/your-life,-pixelated-how-tracking-pixels-watch-your-every-click
  3. Meta Business Help Centre, Help article: restrictions on data sources (including Core Setup). www.facebook.com/business/help/511197658391698
  4. OAIC, Tracking pixels and privacy obligations, 4 November 2024. www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/organisations/tracking-pixels-and-privacy-obligations
  5. Mallesons, Privacy Commissioner zooms in on tracking pixels, 3 July 2026. www.mallesons.com/au/en/insights/latest-thinking/privacy-commissioner-zooms-in-on-tracking-pixels.html
  6. OAIC, Small business. www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/organisations/small-business

Sources were checked on 27 October 2026. Links to third-party sites are provided for reference; Akesis is not affiliated with the organisations listed.