Before I tell any clinic what is wrong with its website, it is fair to show you what was wrong with mine. Every number below comes from my own practice.
Greenland Medical is my private functional medicine practice in Twickenham. An audit in August 2026 found that 88% of its Google traffic was reading about chicken and turkey, Google had declined to index most of its medical pages, and the pages where patients decide to book had no way to book. I rebuilt the site, and enquiries went from around 8 a month to 29 in the first month after relaunch.
Over about twelve months, enquiries to my clinic fell away. I put it down to the cost-of-living crisis, and some of it probably was. Private healthcare is one of the first things people cut.
All that time I was paying for SEO. The reports looked good, and on their own terms they were right: traffic from Google had grown from around 160 clicks a month to more than 2,000. I assumed that meant patients.
By then I had started The AI Growth Clinic to do this work properly for other practices. So in August 2026 I ran the same audit on my own website, using sixteen months of Google Search Console data and a full crawl of the site. It was an uncomfortable diagnosis.
None of this appeared in a monthly report.
| Finding | What it meant |
|---|---|
| Traffic went to the wrong content | Of 17,026 clicks from Google in sixteen months, 88% landed on articles about chicken, turkey and protein. One poultry article drew 2,614 clicks. The Bredesen Protocol page, my flagship service, drew 58. |
| Google declined the medical pages | 31 commercially important pages were crawled and not indexed. They had been written as near-identical keyword variations - ‘functional medicine for brain fog’, ‘…for migraines’, ‘…for chronic fatigue’ - and Google treated them as thin content. |
| No way to book where it mattered | The booking link appeared on 14 pages, but not on the Bredesen page or the fees page: the two pages where patients decide. |
| Enquiries going nowhere | The email link on the team page read ‘greenland-medical.comm’. Anyone who clicked it wrote to an address that does not exist. There was no enquiry form anywhere on the site. |
| No way to see any of it | The site had no working analytics. |
| Hidden technical faults | The sitemap had been failing silently since 2021, with Google finding no pages through it. The Lyme disease page told Google to credit a different page instead of itself. The blog index was titled ‘Social Icons’ and was not in Google at all. |
| Mixed signals about who and where | Three different clinic addresses were in circulation online. The site had no structured data identifying a doctor or a clinic. It described ‘our team’ 81 times, for a practice run by one doctor. |
| Slow, heavy pages | The homepage was 617 KB of code before a single image loaded, with the same logo requested more than 100 times. |
| Links pointing at the wrong page | 302 of 357 links to the site pointed at the homepage, almost all from one regional newspaper syndication network. The Bredesen page had four. |
Sources: Google Search Console (16 months to August 2026), a crawl of the live site on 25-28 August 2026, and the site’s backlink report.
I rebuilt the site from scratch on a fast, static platform and relaunched it on 31 August 2026. The work maps onto the six pillars of the GROWTH Score, the same framework I use for client clinics.
| GROWTH pillar | What changed |
|---|---|
| Ground Truth | One address, used identically everywhere. Written in the first person, as the sole practitioner I am. Structured data identifying the practice as a medical clinic and me as a physician, with my GMC registration. |
| Retrievability | Every one of the 179 addresses Google knew about now resolves, either directly or through a redirect, with none broken. A working sitemap. Pages cut from 617 KB to 23 KB. |
| On-page Answers | The templated keyword pages were consolidated into proper condition pages and long-form patient guides, each opening with a direct answer to the question a patient actually asks. |
| Wider Web Authority | Directory listings being brought into line with the site, starting with the medical directories AI assistants cite most. Future links aimed at the pages that matter, not the homepage. |
| Trust Signals | Credentials, fees and the evidence base set out openly. Clinical claims rewritten to meet UK advertising rules, with no promised outcomes. |
| Head-to-Head Visibility | AI answers for the clinic’s core services now tracked daily against competing clinics, so progress is measured rather than assumed. |
Before the rebuild, the clinic was receiving around 8 enquiries a month. In the first month after relaunch it received 29.
I want to be straight about what that does and does not show. One month is early. Some of the rise may be seasonal, and some comes from fixes as basic as adding a form and a booking button where there were none - which is rather the point. Nothing else about the practice changed in that month: same services, same fees, same doctor. I will update these figures at three and six months, whatever they show.
The lesson I took from it is the one I now give every clinic owner: busy reports are not the same as patients. Traffic had grown twelvefold. It was simply the wrong traffic, arriving at a site that could not turn the right traffic into an enquiry.
Get a free GROWTH Score Snapshot of your clinic, personally reviewed by Dr Andrew Greenland.