Core Web Vitals in 2026: The Numbers That Actually Move Rankings

Every SEO audit mentions Core Web Vitals, and most site owners nod along without knowing which number they're actually chasing. That gap costs rankings. Google doesn't grade on effort. It grades on three specific, published thresholds, measured from real visitors on real devices, and missing them quietly caps how well an otherwise strong page can perform.

The three metrics, plainly

Each metric answers a different question about the experience of loading and using a page. None of them measure how good the content is. All three measure how it feels to sit through.

• Largest Contentful Paint (LCP): how long the biggest visible element takes to render. Google's threshold for 'good' is under 2.5 seconds.
• Interaction to Next Paint (INP): how long the page takes to respond after a click, tap, or keypress. The target is under 200 milliseconds.
• Cumulative Layout Shift (CLS): how much the page jumps around while it loads. Keep it under 0.1.

Field data beats lab data, every time

PageSpeed Insights gives you two scores, and they measure different things. The lab score runs once, on a simulated connection, on Google's own servers. The field score comes from the Chrome User Experience Report: real visitors, on their actual phones, on their actual networks, over the last 28 days. Ranking decisions lean on the field score. A page can post a perfect lab result and still fail Core Web Vitals in Search Console because the people actually visiting it are on a three year old Android phone with patchy 4G, not a fibre connection in a data center.

Where most sites actually lose points

LCP is usually the first metric to fail, and it's rarely the framework's fault. The usual causes are an oversized hero image, a slow time to first byte from the host, and render blocking fonts or scripts sitting in front of the content that should be painting first.

• Serve images in a modern format like AVIF or WebP, sized for where they're actually displayed, not the size they were uploaded at.
• Move to a host or CDN with a fast time to first byte in the regions your visitors are actually in.
• Preload the one hero image or font that's genuinely above the fold, and stop preloading everything else.
• Defer or remove any third party script that isn't earning its place: chat widgets, extra analytics tags, old marketing pixels nobody remembers adding.

INP is a JavaScript problem

A slow response to a tap almost always traces back to one long task hogging the main thread when the click comes in. Big JavaScript bundles that parse and execute all at once are the usual suspect. Splitting code so only what a page needs loads up front, deferring anything non critical, and breaking large tasks into smaller chunks the browser can interrupt all bring INP down without touching the design at all.

None of this requires a rebuild. It requires an audit that looks at the field data first, fixes the one metric actually failing, and leaves the rest of the site alone. That's usually a week of focused work, not a quarter of one.