Page experience: Core Web Vitals, mobile-first indexing and HTTPS
The site-side half is clean: HTTPS, redirects, HSTS, mixed content, resource crawlability and mobile/desktop parity all pass across about 180 URLs. The failures are all in field data, and the trend matters more than the level. Mobile LCP improved sharply, while desktop LCP, mobile INP and CLS all got worse. A single 28-day window would have shown none of that. 5.8 stays Pending: 12 of 25 pages gate content behind a click, and no vetted tool yet tests whether that content is absent from what Google indexes, so the verdict fails toward Pending.
LCP p75 was about 4.4 s on mobile against the 2.5 s threshold, and about 2.8 s on desktop. The breakdown points at resource discovery, not server speed: the largest single contributor was the delay before the LCP image began downloading. The history series tells two stories. Mobile improved 46% from late June, then plateaued at 1.75 times the threshold. Desktop passed in early August, then regressed. Limit: the 28-day window spans at least four production deployments, so the p75 averages different versions of the site. A single window would have shown none of this.
INP p75 was about 450 ms on mobile against the 200 ms threshold, and about 150 ms on desktop. The direction matters more than the level: mobile INP worsened 94% across the history series while the performance work was focused on LCP. The same deployment caveat as 5.1 applies.
CLS p75 was 0.30 on mobile and 0.17 on desktop against a threshold below 0.10. This is the sharpest regression in the cluster: mobile CLS was 0.03 in late June and rose roughly tenfold, and desktop doubled. Lab testing had pointed at one template, but the field data shows the problem is site-wide. The same deployment caveat as 5.1 applies.
The Chrome UX Report returns p75 values over a 28-day real-user window, separately for phone and desktop, as the claim says. The two form factors differ materially. Limit: all 50 URL-level probes returned no data, because Chrome holds too few samples for any single page. Per-page field assessment is unavailable at this site's traffic, which is a property of its volume, not a defect.
Describes how Google's ranking systems weigh page experience. It asserts nothing a site can pass or fail, so it is not measurable.
URL Inspection reports crawledAs: MOBILE for all 25 priority pages, with recent crawl times, so this is current behaviour.
No mobile/desktop differences across a stratified sample of about 180 URLs. Each URL was fetched with both smartphone and desktop Googlebot user agents and compared on title, meta description, meta robots, h1 count, JSON-LD, image alt text and word count.
12 of 25 pages show content behind a 'load more' click in the browser. The claim turns on whether that content is missing from the document Google indexes, and no vetted tool yet makes that comparison, so the verdict stays Pending.
Of about 36,000 subresources referenced across the sample, including scripts, stylesheets, images and iframes, none is disallowed to Googlebot.
All sampled URLs are served over HTTPS, including the few that returned errors. Three transient 502s returned 200 on immediate recheck, so they are recorded as an availability observation.
All sampled http:// URLs return a server-side 301 to the https:// URL. The canonical half is corroborated by 2.2 and 2.3: canonicals are self-referential on https:// pages.
Every sampled response carries Strict-Transport-Security with a one-year max-age and includeSubDomains. The preload directive is absent, but the claim does not require it.
No http:// subresources among about 36,000 on the sampled pages. Limit: this was measured on the Googlebot HTML view, so a subresource injected at runtime by JavaScript would not appear.
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