The gsc-telemetry digest arrived at the same time it always does. Six URLs listed under “URL is unknown to Google.” Same six. Same status. Same as the morning before.
The five legal pages and /stamp-duty-calculator had been live for weeks. The sitemap was submitted. IndexNow had been pinged. Twelve manual Request-Indexing nudges had gone through Search Console. Every operator-side lever pulled. The pages were still unknown to Google.
PR #153 on 24 June was my code review of those six surfaces. The verdict: no technical SEO blocker across any of them. HTTP 200, index, follow on the robots directive, canonical pointing at self, prerender unblocked after a fix in the same commit. I cleaned up the FAQ structured data. I simplified the legal layout. By the time the review closed, nothing left on the technical side could explain why Google hadn’t found them.
The pages weren’t unindexable. They were unfindable.
That distinction is the article.
What the code review ruled out
Improving on the prior coverage of these six pages: that article established what I found and ruled out. This piece is about what the problem actually was.
Every technical signal on the six surfaces was correct after PR #153:
- HTTP 200 on all six URLs
- Robots directive:
index, follow - Canonical: self-referencing on each
- Prerender: unblocked
- FAQ structured data: valid
- Legal layout: rendering noise removed
After those fixes shipped, nothing remained for a crawl audit to flag. Indexable, by every measurable technical signal. And yet the gsc-telemetry digest continued to list them under “unknown.”
Why manifests aren’t enough
A sitemap is a manifest. It tells Googlebot “these URLs exist; please crawl them.” For high-trust domains with established crawl frequency, that request is serviced promptly. For newer domains, for pages outside the main content hierarchy, or for URLs that haven’t yet accumulated inbound crawl-weight, the manifest entry joins a queue. How fast that queue moves depends on how much Googlebot trusts the site overall. If the trust is low, the entry waits.
IndexNow and manual Request-Indexing submissions work similarly. They’re direct channels to Google’s indexing queue. Worth doing: the 20 June operator-side push was the right first move. They establish baseline hygiene and rule out technical causes. But they don’t accelerate queue servicing for pages that haven’t cleared Google’s trust threshold. They just ensure the entry is in the queue.
Editorial links work differently.
When a page Googlebot already trusts and regularly crawls contains a link to another page, Googlebot follows it. It’s not working through a manifest. It’s following a path it discovered in content it was already reading. The destination page receives crawl credit from the source. That credit is weighted by context: body prose carries more weight than a repeated footer entry.
This is not a subtle distinction in Googlebot’s crawl algorithm. Sitemaps and footer links both tell Googlebot a URL exists. An editorial body link tells Googlebot that a trusted author directed readers there, mid-sentence, in content that already earns crawl attention. That signal is categorically harder to fabricate at scale, which is why it carries more weight.
The specific problem
The five legal pages and /stamp-duty-calculator were footer-only pages. Every internal link to them, prior to 24 June, was in the site footer. Footer links are structurally low-weight: Googlebot knows footers repeat across every page on a site, so the crawl-weight assigned to a footer-linked URL is fractional. The first time a footer link appears, it carries some weight. After it has appeared on several hundred pages, the marginal weight of each additional instance is negligible.
For a page trying to move out of “unknown to Google,” fractional repeated footer-weight is not sufficient. The queue entry sits waiting for a crawl justification that fractional weight doesn’t provide.
The fix
The decision on 24 June was to add body-prose internal links from six published articles directly to /disclaimer and /stamp-duty-calculator. Not new footer links — the footer already had them. Editorial prose links, in context, in the body of content Googlebot was already crawling because those articles had traffic and inbound links of their own.
The devlog describes this as “crawl-weight transfer to footer-only pages.” That phrasing is accurate. The six articles already in Googlebot’s crawl path were the mechanism. Each body link in an article about stamp duty or tax liability, placed in a sentence about the limitations of automated estimates, transfers crawl-weight from a trusted source to a page that needs it.
The links were written to be genuinely contextual. A disclaimer link in the body of a tax calculator article, in a sentence directing readers to check the site’s legal disclosures before relying on estimates, carries both crawl-weight and editorial justification. It is not a manufactured signal.
Out of scope: third-party backlinks, RSS pings, social signals. The intervention was internal. The asset base was what it was.
The signal chain
gsc-telemetry made this diagnostic path visible. Built during Phase 1.6 as a daily skill, it surfaces URL-level status from Search Console without a manual log-in each morning. Same six URLs, same status, every day: a pattern legible only because the cadence was automated. Weekly manual checks would surface the same information, but less clearly. Persistence is what makes “unknown” a signal rather than noise, and daily resolution is what makes persistence visible.
The sequence that produced the fix:
- gsc-telemetry digest surfaces the six stuck URLs daily
- Persistence across multiple weeks signals that standard operator-side tooling has been exhausted
- I rule out every technical cause via the PR #153 code review
- Technical dead-end → editorial diagnosis
- Body-prose links added from six existing articles to the two destinations
Whether the “unknown” status clears after step 5 is what subsequent gsc-telemetry digests will show.
What this isn’t
It isn’t a claim that sitemaps are useless. The baseline hygiene matters. Every URL on the site should be in the sitemap; the sitemap should be error-free. The 20 June canonical fix was a genuine improvement.
It’s also not a claim that editorial links always outperform IndexNow. For a well-established domain with strong crawl frequency, IndexNow works. The argument is narrower: for pages that haven’t cleared Google’s trust threshold on a lower-authority site, manifest-based approaches queue a request that may wait indefinitely. Editorial links provide a discovery path Googlebot follows because it trusts the source, independent of the queue.
Footer-only pages on a newer site are in that position. Editorial prose is the right tool for that configuration.
The broader principle
When the operator-side push is exhausted and the telemetry signal persists, the next question isn’t “what other technical signal can I send?” It’s “which pages already in my crawl path can transfer weight to the stuck pages?”
That’s an editorial question, not a technical one. The answer is in the content that already ranks.
The gsc-telemetry skill is what made the pattern visible quickly enough to diagnose correctly. Without the daily digest, the technical audit would have happened eventually, the editorial intervention would have been considered eventually. The telemetry compressed the diagnosis cycle. That’s what it’s there for.
