Real users · qualified evidence · privacy boundaries
Field performance is published only when the sample can support it.
This page reports first-party LCP, INP and CLS at the 75th percentile over a rolling 28-day window. A mobile or desktop row remains unpublished until all three metrics have at least 75 eligible observations.
0/18 established rowsCurrent qualified coverage across nine sites and two device groups.
28-day window2026-07-09 through 2026-08-05, UTC.
No visitor profileOnly aggregate histograms are retained.
“Awaiting traffic” and “Collecting” deliberately expose no percentile. Synthetic visits, audits and automated browsers are excluded and must never be used to manufacture eligibility.
Publication threshold
A row becomes established only when LCP, INP and CLS each have at least 75 eligible observations in the same rolling window. INP requires a real interaction, so it commonly qualifies last.
Good thresholds at p75 are LCP ≤2,500 ms, INP ≤200 ms and CLS ≤0.10. Values above 4,000 ms, 500 ms and 0.25 respectively are poor.
Machine-readable evidence
The same qualified rows are available as versioned JSON. Low-volume rows disclose status but no metric value or sample count.
The browser client uses the official GoogleChrome/web-vitals library. It stores no cookie, persistent visitor identifier, IP address, query string, full referrer URL, user-agent string or raw measurement. Do Not Track, Global Privacy Control, known bots, automated browsers, private application routes and audit visits are excluded.
This is first-party field evidence. It is not Chrome UX Report, Search Console or Lighthouse data. The public PageSpeed field panel returned “No Data” for all nine homepages on 5 August 2026; that means insufficient public CrUX evidence, not a good or poor result. None of these metrics proves ranking, product adoption, editorial endorsement or AI citation.