The three metrics
- LCP (Largest Contentful Paint) — how fast the biggest thing on screen loads. Target under 2.5s.
- INP (Interaction to Next Paint) — how snappy your site feels when clicked or tapped. Target under 200ms.
- CLS (Cumulative Layout Shift) — how much the page jumps around as it loads. Target under 0.1.
How much do they affect rankings?
Honestly? Less than people claim. Core Web Vitals are a tiebreaker, not a primary ranking factor. A faster site rarely outranks a more relevant slow one. But when you're close to a competitor on relevance, passing CWV tips the scale — and the conversion-rate gains alone usually justify the work.
What to fix first (in order)
- Compress and serve images properly. WebP/AVIF, correct dimensions, lazy-loaded below the fold. This single fix solves most LCP problems.
- Reserve space for images and ads. Always set width/height attributes. Kills CLS.
- Cut JavaScript. Remove abandoned tracking pixels, chat widgets you don't use, and old A/B tools. Biggest INP win for most sites.
- Use a CDN. Cloudflare's free tier alone often shaves 500ms off LCP.
- Preload the hero image.
<link rel="preload" as="image">for above-the-fold imagery.
How to measure (properly)
- Use the Core Web Vitals report in Google Search Console — it shows real-user data, which is what Google actually uses.
- PageSpeed Insights for ad-hoc checks. Pay attention to the "Origin Summary" — that's the 28-day field data.
- Don't chase the lab score. A 100 in Lighthouse with failing field data is worse than an 85 with passing field data.
What to ignore
Server response time micro-optimisations under 200ms. Render-blocking warnings on tiny CSS files. Third-party "speed plugins" that promise miracles. Fix images, fix layout shift, cut JavaScript — that's 90% of the win.