Indexing, canonicalization, HTTP status, redirects and URL structure
Canonicalization is the strong part: Google's chosen canonical matches the site's declaration on all 25 priority pages, and every redirect is a server-side 301. The two Partials are 2.8 (soft-404 exposure, mostly fixed on re-measurement) and 2.9 (a small set of badly encoded URLs out of everything the sitemaps declare).
Google's chosen canonical matches the site's declared canonical on all 25 priority pages, and all 25 are 'Submitted and indexed'. An independent read of the Googlebot HTML agrees 25 of 25, so two tools on two views give the same answer. Limit: URL Inspection reports Google's last-indexed state, so this is Google's view as of its own last crawl, not a live fetch.
Every priority page serves exactly one <link rel=canonical>, inside <head>, with an absolute self-referential URL (trailing slash included in the comparison). Measured on the Googlebot HTML view.
All three signals name the same URL on all 25 priority pages: 200 with no redirect, a self-referential canonical, and membership in the live sitemaps. Across about 1,300 declared URLs, none redirects and none carries a canonical naming a different URL. Five declared URLs return 404, but that is a sitemap defect, so it carries its own Gap under 1.8.2.
Both delivery channels are in use with exact syntax. The 404 template serves a robots meta tag with noindex, nofollow in the documented comma form, and internal-search URLs serve X-Robots-Tag: noindex. None of the 25 indexable priority pages carries noindex in either channel.
The one place a noindex meets a robots.txt disallow is internal search. The trap was avoided in the documented order: those URLs were confirmed 'Excluded by noindex' in Search Console before the Disallow was added. Residual risk, stated plainly: while the Disallow stands, Googlebot cannot see the noindex on any new search URL, so a newly linked one could still be indexed URL-only.
Every redirect found is permanent. 65 URLs were probed with redirects not followed, so each hop was read on its own: 62 single-hop 301s, one 301 to 301 to 200 chain, and 2 plain 200s. No 302, 303 or 307 anywhere, and every chain ends on a 200.
All redirects are server-side, with no meta refresh and no JavaScript location redirects in the Googlebot HTML. Every redirect fronts a GET navigation, where 301 is correct, so the claim's preference for 308 has nothing to apply to. That is reported, not raised as a defect.
The 4xx half holds: 6 of 7 deliberately nonexistent URLs return a genuine 404, and the 404 template is itself noindexed. On the 2xx half, 91 declared URLs had served placeholder or thin content on a 200. On re-probing those 91 after the fix, 86 pass and 5 do not, and median words in <main> went from 174 to 1,281. Only those 91 were re-probed, so this means the known failures are mostly fixed, not that the whole site was re-swept. Kept at Partial because 5xx behaviour is untested (we will not induce it on production), and any path under a valid page returns 200 with the parent page and a canonical to the parent.
Stayed Partial, but the failing half shrank: of the 91 URLs that served placeholder or thin content, 86 now pass. Recorded as a correction rather than a silent re-tally, because the earlier number no longer described the site.
Every URL the sitemaps declare was scanned, not a sample: no underscores, query strings, raw non-ASCII characters, spaces, double slashes or trailing slashes, and one host and one scheme. 36 URLs miss the bar. 14 carry percent-escapes that should never reach a slug, including destroyed accents, non-breaking spaces and encoded pipes, and 23 break the site's lowercase casing. Impact is low: 3 of the 36 registered any impressions in 28 days, with 0 clicks, and no priority page is affected. The recommendation targets the slug generator, not a bulk rename.
Want this run on your site?
We test your site against the same claims, record a verdict and the evidence for each one, and tell you what to fix first.
Start a conversation