The morning gsc-telemetry digest on 24 June flagged six surfaces as “URL is unknown to Google” — the same six that had been in that state since Phase 1.6 began. Five legal pages. /stamp-duty-calculator. All returning HTTP 200. All set index, follow. Canonical pointing to themselves. Nothing obviously wrong.
The obvious move was a thorough audit. PR #153 went through ultracode — the multi-agent review harness that fans out across a diff and reports findings by dimension. It came back clean on every technical SEO category: HTTP status, meta robots, canonical chain, prerender behaviour, schema, FAQ structure, layout. The pages could be indexed. They just weren’t being found.
Improving on the earlier diagnostic work documented alongside this sprint: the gsc-telemetry tooling had established that these six pages had no technical reason to be invisible. This article covers what the actual fix turned out to be, and what the diagnostic gap was.
What the ultracode review did find
The review wasn’t empty. Three categories of fixable issues surfaced: a prerender blockage on one of the legal pages, a FAQ helper doing structural work it shouldn’t have been, and layout inconsistencies across the legal surface. All three landed in the fix(seo) PR.
None were crawl blockers. A prerender issue affects the quality of what Googlebot sees when it arrives, not whether it arrives. Schema correctness matters after the first visit. The pages were cleaner after the PR. Googlebot still hadn’t visited.
What “URL is unknown to Google” actually means
Google distinguishes between “crawled but not indexed” and “URL is unknown to Google.” The first means Googlebot visited and made a decision. The second means Googlebot hasn’t visited — the URL isn’t in its crawl frontier.
A page outside Googlebot’s frontier has no inbound link equity from a crawled surface. Technical correctness is irrelevant. Schema doesn’t factor in. Sitemap submissions help at the margin but don’t substitute for real crawl paths.
The six pages had one thing in common: their only inbound links came from the site footer. Every page on the site linked to them via footer nav. That’s a pattern Googlebot reads as structural chrome. Footer links repeat identically across hundreds of pages — Googlebot assigns them low crawl weight because they’re navigation, not editorial endorsements. The pages were reachable. They weren’t being reached.
Why the ultracode review couldn’t catch this
The ultracode review checked the pages themselves. That’s the right scope — the question being asked was whether the pages had technical blockers. They didn’t. But “why isn’t Googlebot finding these pages?” is a different question. It’s not about the pages. It’s about the link graph.
A page-level technical audit of six clean pages returns clean. The finding “these pages are link-starved” requires looking at inbound links from other pages — outside the diff scope of a page-level review.
Out of scope for a page-level technical SEO audit: cross-page link graph position. Worth adding as an explicit diagnostic question at the start of any future “URL is unknown to Google” investigation: not just “is this page technically sound?” but “where does it sit in the crawl graph?”
The fix
Once the diagnosis landed, the fix was mechanical. Six existing articles had natural editorial space for a /disclaimer or /stamp-duty-calculator link. One contextual sentence per article. One in-body link. Six changes, one chore(content) PR.
Selection logic:
/disclaimerlinked from articles that reference data sources, estimate accuracy, or editorial policy on financial figures. The link serves the reader — the article has a genuine reason to point there./stamp-duty-calculatorlinked from articles covering stamp duty thresholds, property purchase costs, or property-related tax.
No new pages. No structural changes. No sitemap amendments.
What this actually transfers
A footer link and a body-prose link carry the same href. The HTML is equivalent. The crawl signal is not.
A footer link appears on every page, in the same structural position, as site-wide chrome. Googlebot knows this pattern. It assigns crawl priority proportional to estimated editorial intent. A link identical across hundreds of pages is structural. Crawl weight: low.
A body-prose link appears in one paragraph, in one article, about a related topic. The editorial intent is legible from context. The crawler treats it as a content recommendation — one piece of content pointing to another as a related resource. Crawl weight: proportional to the linking page’s authority and topical relevance.
Same destination URL. Different graph position. Different treatment.
The right diagnostic question for “unknown to Google”
The distinction between the two Search Console statuses determines the fix.
“Crawled but not indexed”: Googlebot visited, made a decision against indexing. Root cause is content quality or thin content. Fix: improve editorial depth.
“URL is unknown to Google”: Googlebot hasn’t visited. Root cause is link graph position. Fix: add in-body links from crawled pages.
These are different remediation chains. Treating them as the same wastes effort. The 24 June morning was spent confirming the pages were technically receivable. The afternoon was the actual fix. Separating these two diagnostic questions at the start would have made the morning a much shorter step.
What comes next
The chore(content) commit is in. Whether Googlebot picks up the six pages on the next cycle isn’t observable yet — Search Console coverage updates on its own schedule. The signal to watch: “URL is unknown to Google” transitioning to any of the crawled or indexed states in the fortnight following the link additions.
Either transition confirms the links did their job. What Googlebot decides after it arrives is a content-quality question, and the fix(seo) PR already handled that side.
The gsc-telemetry skill surfaces the daily digest. No manual monitoring required.
