CWV that affects rankings is field data — the Chrome User Experience Report (CrUX) data Google collects from real Chrome users. Lab data (Lighthouse, PageSpeed Insights' lab section) is useful for diagnosis but is not what Google ranks on.
If your CrUX data is missing (low traffic), Google falls back to algorithmic estimates from your origin's other pages and similar sites. Either way: optimize for real-user performance, not for a synthetic Lighthouse score.
Modest but real. CWV is a tiebreaker — between two pages with similar relevance and authority, the faster one wins. It is rarely the difference between page 3 and page 1, but it is often the difference between position 4 and position 2 in tightly contested SERPs.
The bigger payoff is conversion rate. Bringing an LCP from 4.5s to 1.8s typically lifts conversion rates 10-25% — which dwarfs the SEO impact for most businesses.
INP (Interaction to Next Paint) replaced FID in March 2024. INP measures the responsiveness of all interactions during the page session, not just the first one — which is a much better proxy for real user-perceived responsiveness.
No — CWV is measured separately for mobile and desktop, and Google primarily uses the mobile data for ranking on mobile-first-indexed sites (which is essentially all sites in 2026).
CrUX uses a 28-day rolling window — meaningful improvements typically appear in CrUX 2-4 weeks after deployment.
No. A 100 Lighthouse score is a synthetic-test artifact and frequently does not match real-user performance. Optimize for CrUX field data; treat Lighthouse as a diagnostic tool only.