# LoveIslands.uk: SEO baseline and implementation

Audit date: 9 November 2026. Self-initiated project, standard depth.

The baseline CSV is the actual Google Search Console Web performance export
for 8 August–7 November 2026: 92 daily rows, 104 clicks, 13,396 impressions.
21–27 October totals 753 impressions and 6 clicks. 28 October–3 November
totals 77 impressions and 2 clicks. Impressions declined 89.77%. This is a
comparison of two seven-day periods; it is not an explanation of the decline.

Sources: authorised, logged-in Search Console and Analytics reports; public
HTTP responses; rendered DOM; three mobile PageSpeed Insights tests.
Collection and analysis were assisted by Codex. The project owner should
review the implementation and evidence before claiming personal outcomes.

The initial crawl used a synthetic Googlebot user agent, JavaScript off,
concurrency 2 and a 350-request limit. It was not a verified Googlebot IP or
an exhaustive crawl. All 214 current sitemap URLs returned 200 in that pass.
The historic 404 list and current sitemap are separate cohorts. Search
Console query and page tables use different aggregations and must not be
joined as if they contain verified query-by-page pairs.

Full source and account screenshots are not included in the public evidence.
Repository access may require the project owner's permission. Verify HTTP
behaviour with scripts/seo-http-checks.cjs against a built local Worker and
initialised local D1. The expected checks cover host redirects, pagination
canonical/title, first-page redirects, real invalid-page 404s, unavailable
archives, single article H1 and the public project evidence path.

Analytics loads after permission and uses G-4GB3JZQ39Y. Manual page_view
events use send_page_view:false. This stream's enhanced measurement was
turned off on 9 November to avoid duplicate history views and automatic
search/form data. Verify DebugView before calling collection fully accepted.
Declined visitors and blockers
will be absent from Analytics; link events do not prove completed reviews.

No search recovery, ranking improvement, recruitment result or real-user
Core Web Vitals pass has been established. Source code identifies a major
29 October relaunch, but timing alone does not establish causation. Worker
CPU-limit errors observed during the audit need production logs to assess.