Core Web Vitals

Core Web Vitals optimization

Targeted optimization for the three Core Web Vitals using field data to identify user impact and browser traces to diagnose causes.

Measure before optimizing

When a runnable URL is available, read the performance measurement workflow (opens in a new tab). Prefer this sequence:

1. Check page-level CrUX p75 data, with a clearly labeled origin fallback when page data is unavailable.
2. Record a browser performance trace under stated conditions. With Chrome DevTools MCP, trace summaries can include CrUX alongside the observed lab metrics.
3. Analyze only the insights associated with the failing metric, then inspect the implicated code and resources.
4. Re-run equivalent lab measurements after the fix. Do not claim an immediate field improvement; CrUX and first-party RUM need new user visits.

If only source code is available, identify likely causes but do not claim that LCP, INP, or CLS is failing without runtime evidence.

The three metrics

Google measures at the 75th percentile — 75% of page visits must meet “Good” thresholds.

LCP: Largest Contentful Paint

LCP measures when the largest visible content element renders. Usually this is:

• Hero image or video
• Large text block
• Background image
• <svg> element

Common LCP issues

1. Slow server response (TTFB > 800ms)

Fix: CDN, caching, optimized backend, edge rendering

2. Render-blocking resources

<!-- ❌ Blocks rendering -->
<link rel="stylesheet" href="/all-styles.css">

<!-- ✅ Critical CSS inlined, rest deferred -->
<style>/* Critical above-fold CSS */</style>
<link rel="preload" href="/styles.css" as="style" 
      onload="this.onload=null;this.rel='stylesheet'">

3. Slow resource load times

<!-- ❌ LCP image is discovered only after a stylesheet loads -->
<div class="hero"></div>

<!-- ✅ Discoverable in initial HTML and prioritized -->
<link rel="preload" href="/hero.webp" as="image" fetchpriority="high">
<img src="/hero.webp" alt="Hero" fetchpriority="high">

Prefer a discoverable <img> with fetchpriority="high". Add the preload only when the trace shows that the resource would otherwise be discovered late; duplicate or speculative preloads can compete for bandwidth.

4. Client-side rendering delays

// ❌ Content loads after JavaScript
useEffect(() => {
  fetch('/api/hero-text').then(r => r.json()).then(setHeroText);
}, []);

// ✅ Server-side or static rendering
// Use SSR, SSG, or streaming to send HTML with content
export async function getServerSideProps() {
  const heroText = await fetchHeroText();
  return { props: { heroText } };
}

5. Make navigations instant with the Speculation Rules API

For sites with predictable same-origin journeys, prerendering a likely next page can make a successful subsequent navigation much faster. Treat this as a measured navigation optimization, not a substitute for fixing the current page’s LCP.

<script type="speculationrules">
{
  "prerender": [{
    "where": { "href_matches": "/*" },
    "eagerness": "moderate"
  }]
}
</script>

Current Chrome behavior is specific enough to guide the choice:

Start conservatively and measure prediction hit rate, transferred bytes, server load, and navigation improvement before expanding the rules. Recheck Chrome’s maintained eagerness documentation (opens in a new tab) before hardcoding timing-sensitive behavior.

Caveats:

• Bandwidth/CPU cost. Each prerender is roughly a full page load. Scope where carefully (href_matches patterns, exclude logout/checkout) and avoid immediate outside small sites.
• Side effects fire early. Analytics, ads, and any code that runs on load will fire when the prerender starts, not when the user navigates. Gate side effects on the prerenderingchange event (opens in a new tab) or document.prerendering.
• Chromium-only. Safari and Firefox ignore the script — it’s a progressive enhancement, never a regression.

LCP optimization checklist

- [ ] TTFB < 800ms (use CDN, edge caching)
- [ ] LCP resource is discoverable in initial HTML and prioritized; preload only if the trace shows late discovery
- [ ] LCP image optimized (WebP/AVIF, correct size)
- [ ] Critical CSS inlined (< 14KB)
- [ ] No render-blocking JavaScript in <head>
- [ ] Fonts don't block text rendering (font-display: swap)
- [ ] LCP element in initial HTML (not JS-rendered)
- [ ] Speculation Rules added for likely-next navigations (moderate eagerness)

LCP element identification

This snippet diagnoses the current page session. It is not field data.

// Find your LCP element
new PerformanceObserver((list) => {
  const entries = list.getEntries();
  const lastEntry = entries[entries.length - 1];
  console.log('LCP element:', lastEntry.element);
  console.log('LCP time:', lastEntry.startTime);
}).observe({ type: 'largest-contentful-paint', buffered: true });

INP: Interaction to Next Paint

INP measures responsiveness across clicks, taps, and key presses during a visit. Diagnose its input delay, processing time, and presentation delay separately; a slow interaction may involve main-thread contention before the handler, expensive application work, or delayed rendering after it.

When field INP is poor or a trace identifies a slow interaction, read the INP reference (opens in a new tab) for trace interpretation, yielding patterns, third-party and rendering causes, a single-session observer, and first-party attribution.

CLS: Cumulative Layout Shift

CLS measures unexpected layout shifts across a page visit. Use field attribution or a trace to identify the shifted node and the trigger; do not assume the visible victim caused the shift.

When field CLS is poor or a trace reports shifts, read the CLS reference (opens in a new tab) for reserved-space patterns, dynamic content, font and animation fixes, a debugging observer, and a verification checklist.

Measurement sources

Do not route performance through Chrome DevTools MCP’s lighthouse_audit; that capability intentionally covers non-performance Lighthouse categories. Do not compare a single lab value directly with a field p75 as if they were equivalent samples.

When adding or reviewing production collection, read the first-party RUM reference (opens in a new tab). Prefer the web-vitals library because raw browser APIs do not by themselves implement every Core Web Vital’s lifecycle and reporting rules.

Framework quick fixes

Next.js

// LCP: Use next/image with priority
import Image from 'next/image';
<Image src="/hero.jpg" priority fill alt="Hero" />

// INP: Use dynamic imports
const HeavyComponent = dynamic(() => import('./Heavy'), { ssr: false });

// CLS: Image component handles dimensions automatically

React

// LCP: Preload in head
<link rel="preload" href="/hero.jpg" as="image" fetchpriority="high" />

// INP: Memoize and useTransition
const [isPending, startTransition] = useTransition();
startTransition(() => setExpensiveState(newValue));

// CLS: Always specify dimensions in img tags

Vue/Nuxt

<!-- LCP: Use nuxt/image with preload -->
<NuxtImg src="/hero.jpg" preload loading="eager" />

<!-- INP: Use async components -->
<component :is="() => import('./Heavy.vue')" />

<!-- CLS: Use aspect-ratio CSS -->
<img :style="{ aspectRatio: '16/9' }" />

References

• Detailed LCP optimization (opens in a new tab) — read when an LCP trace points to discovery, loading, or render delay
• Detailed INP optimization (opens in a new tab) — read when a trace or field attribution identifies a slow interaction
• Detailed CLS optimization (opens in a new tab) — read when a trace or field attribution identifies unexpected shifts
• web.dev LCP (opens in a new tab)
• web.dev INP (opens in a new tab)
• web.dev CLS (opens in a new tab)
• Performance skill