Scope and measurement window
We compared Google Search Console UK Web data for 27 August to 23 September 2026 with 30 July to 26 August 2026. Both windows contain 28 days. The later window includes time before and after the launch, so it is not four complete weeks of post-launch performance.
The comparison was read on 26 September. We retain the source observations and release evidence in our internal audit record. Search reporting can change as data settles; this page records that dated observation.
The measured search result
UK clicks rose from 76 to 110, an increase of 44.7%. UK impressions fell from 37,040 to 35,537, a decrease of 4.1%. Average position improved from 21.2 to 19.8. These measures describe Google Search visibility and visits from search results, not qualified leads or revenue.
The shorter post-launch comparison was also mixed: 8 to 21 September recorded 59 UK clicks versus 42 for 25 August to 7 September, while impressions fell by approximately 25%. Changing query mix, competition and other factors can affect these totals. The observations do not isolate the effect of the redesign.
Important commercial searches did not all improve
For the same 28-day UK comparison, dental marketing agency moved from average position 8.8 to 10.5 and retained one click in each period. Aesthetic marketing agency moved from 6.5 to 11.3, with clicks falling from two to zero. SEO for clinics improved from 4.7 to 4.4 and gained two clicks from zero.
A domain-wide improvement can therefore hide weaker commercial searches. We review those pages and queries separately, while accounting for the old and replacement URLs during migration.
What we corrected after the review
The verified release work repaired three valuable legacy redirects, improved two treatment pages and their discovery links, removed duplicate breadcrumb markup on nine sector pages, revised two substantial guides and clarified buyer questions on two sector pages. Forms now suppress explicitly identified internal QA submissions from future lead events.
The second release used matched hosting files and database recovery points. A local 194-page check examined 13,210 links without finding a broken internal route. Public verification matched 340 published files and passed 21 adjacent checks. Those are deployment checks, not Google indexing or lead-growth evidence.
Why we are not claiming more customers
Three reviewed website receipts comprised two internal QA tests and a supplier solicitation. Those records do not establish qualified demand. They also do not prove that the entire business received no genuine enquiries through other routes.
The CRM’s attribution view had a workspace lead total but no source rows and incomplete booking and revenue links. We therefore cannot attribute customer growth to this release. The remaining reporting work must connect genuine enquiries, booked calls and customer records, and keep unknown outcomes visible.
Performance and AI-search limits
A mobile Lighthouse test of the dental sector page on 26 September reported performance 98, accessibility 100, best practices 100 and SEO 100. Largest Contentful Paint was 2.4 seconds, Total Blocking Time 20 milliseconds and Cumulative Layout Shift 0.001. It was one lab run before the final FAQ addition, with no field data available.
Google’s AI-search documentation says the usual SEO foundations apply. Clear answers and accurate supporting evidence can improve usefulness, but no special schema guarantees an AI citation. We have not demonstrated an increase in AI-generated referrals or customers from these changes.
What another clinic can learn
Keep an old-URL inventory, preserve relevant destinations, test the actual contact journey and compare equal reporting windows. Separate technical acceptance from indexing, visibility from enquiries, and enquiries from customers. Publish both favourable and unfavourable observations with the definitions needed to interpret them.
Explore our SEO guide and evidence hub for the wider approach. The results on this page are specific to ClinicGrower’s own website and should not be used as a forecast for another business.