A new website with no drop in visits from Google, and a booking form that counts its requests

The practice works with its clients one at a time. Its guides had built a strong position in Google over the year before, and clients found the practice through them. The website itself ran on a hosted website builder.

What was wrong

Most of the old site was not the site. Of about 230 pages Google could reach, only about 54 held real content; the rest were archive pages the website builder made on its own, and 26 internal links led to pages that no longer existed. The guides ranked, but none of the pages describing the service did, so a reader could read a guide and never meet the service. And nobody could say how many people asked for a consultation, because no one counted.

What I did

• A new website, built in Astro. Every page that earned visits kept its address or forwards to its new one, the archive pages and dead links went, and the site tells search engines who runs the practice and what it offers. It went live in August.
• Guides that lead to the service. The guides that bring in the most visits now introduce the service where it helps the reader, and link straight to the service page.
• A booking form that counts. Every consultation request is now recorded, every enquiry gets a personal acknowledgement, and the form asks how they heard about the practice, so requests can be credited to their source for the first time.
• Guides kept current. Guides were brought up to date, each checked on its primary source, and pages that promised things the practice does not offer were corrected.

What changed

	Before	After
Visits from Google, September	185 (2025)	527 (2026)
Visits from Google per full week, September	34 to 50 (2025)	114 to 139 (2026)
Consultation requests, counted	not counted	21 in the first six weeks

Visits from Google did not drop in the move. Every full week since the new site went live has beaten every week of July. September brought the most visits from Google of any month since January, and about 2.8 times September a year earlier. Part of that comparison reflects the year of growth before the move, so the fair reading is this: the site came out of the move far stronger than it stood a year earlier.

One in three visitors who open the booking page go on to send a request: 18 of 54 in the first six weeks, counting the requests Analytics could track. Of the 18 requests whose source could be tracked, 5 came through AI assistants such as ChatGPT, from under 3 percent of visits. In October, 14 more requests arrived in under three weeks.

What has not moved yet

The guides bring in far more readers than the service page. In the first six weeks, the strongest guide had about 250 readers while the page describing the service had 16. The links from the guides to the service are the first step; the next is giving the service pages the words people actually search for.

How we measured

• Visits from Google: Google Search Console, clicks, by month and by week, September 2025 against September 2026, read October 2026. Appearances are not compared across that year, because Google miscounted them for part of it; clicks were counted correctly throughout.
• No loss in the move: Google Search Console, clicks per full week, July 2026 against every full week from the launch in August to early October 2026.
• Consultation requests: the booking system’s own records, late August to early October 2026, and 14 more in October 2026.
• Booking page visitors who sent a request, and the guide and service page readers: Google Analytics 4, late August to early October 2026.
• Requests from AI assistants: Google Analytics 4 and the booking system, the 18 requests whose source could be tracked, late August to early October 2026.
• Pages on the old site: a full crawl of the old site before the move, 2026.