Core Web Vitals: The Three Metrics Google Actually Uses for Rankings
In May 2021, Google rolled out Core Web Vitals as an official ranking factor. Unlike most algorithm updates that are vague or difficult to measure directly, this one gave us specific, measurable targets. Three metrics, clear thresholds, data available in free tools.
I've spent the last few years optimizing sites specifically around these metrics. Here's what actually matters and what you can do about it.
The Three Metrics
LCP — Largest Contentful Paint
LCP measures how long it takes for the largest visible element on the page to load. Typically this is a hero image, a large heading, or a featured image in a blog post.
- Good: under 2.5 seconds
- Needs improvement: 2.5–4 seconds
- Poor: over 4 seconds
LCP is the most impactful metric for most sites because it most closely correlates with "does this page feel fast?"
INP — Interaction to Next Paint (replaced FID in March 2024)
INP measures how quickly your page responds to user interactions — clicks, taps, keyboard input. It's measured across all interactions during a page visit, not just the first one.
- Good: under 200ms
- Needs improvement: 200–500ms
- Poor: over 500ms
INP issues are usually caused by heavy JavaScript that blocks the browser's main thread.
CLS — Cumulative Layout Shift
CLS measures how much the page layout shifts unexpectedly while loading. You've experienced bad CLS when you go to click a button and an ad loads above it, moving the button just as your finger touches the screen.
- Good: under 0.1
- Needs improvement: 0.1–0.25
- Poor: over 0.25
CLS is often the easiest to improve and the most annoying to users when left unaddressed.
Why These Three
Google designed Core Web Vitals to measure user experience dimensions that correlate with satisfaction and engagement. Loading performance (LCP), interactivity (INP), and visual stability (CLS) were chosen based on large-scale user data showing they predict whether users complete tasks on a page.
The research behind CWV comes from Google's analysis of millions of page loads and the correlation between these metrics and user abandonment rates. That's why they're ranking factors — they proxy for whether users actually have a good experience.
How to Diagnose Your Current State
Google Search Console — Under Experience > Core Web Vitals, you'll see which pages have Good, Needs Improvement, or Poor status. This uses real-world user data (Chrome User Experience Report), which makes it the most meaningful measure.
PageSpeed Insights — Shows both field data (real users) and lab data (simulated test). The field data section shows your Core Web Vitals from actual visitors. The lab data section is useful for diagnosing issues even on pages with limited real-world data.
Chrome DevTools — Under Performance, you can record page loads and see exactly which operations are causing delays. More technical, but necessary for diagnosing complex INP issues.
Fixing LCP
The most common LCP issues:
Slow server response. If your server takes 600ms to return the first byte of HTML, you've already used 600ms of your LCP budget before any content loads. Solutions: caching, CDN, better hosting.
Unoptimized images. A 2MB hero image in JPEG format will fail LCP on most connections. Convert to WebP, compress, and use the loading="eager" attribute on the LCP image (override lazy loading for the above-fold image).
Render-blocking resources. CSS and JavaScript in the <head> that blocks HTML rendering delays LCP. Defer non-critical scripts; inline critical CSS.
Slow third-party resources. Analytics, chat widgets, and ad scripts loaded in the header can delay LCP significantly. Move them to load after the main content.
Fixing INP
INP failures are almost always JavaScript-related.
Long tasks on the main thread. The browser's main thread handles both JavaScript execution and rendering. When JavaScript runs a long task (over 50ms), user interactions are queued until it finishes. Break long tasks into smaller chunks using setTimeout or scheduler.yield().
Third-party scripts. Marketing pixels, analytics, A/B testing tools, chat widgets — these often run JavaScript on the main thread. Load them with async or defer. Audit which ones you actually need; unnecessary third-party scripts are the most common INP problem on marketing sites.
Heavy event listeners. Event handlers that do too much work synchronously on user interaction will cause INP failures. Profile interactions in Chrome DevTools to identify which handlers are slow.
Fixing CLS
CLS issues are usually caused by content without reserved dimensions.
Images without width/height attributes. When the browser doesn't know an image's dimensions before loading it, it renders without space, then shifts the layout when the image loads. Fix: always include width and height attributes on <img> tags, or use aspect-ratio in CSS.
Ads without reserved space. Ad slots that expand when an ad loads cause significant CLS. Reserve space for ads with explicit dimensions even when no ad is loaded.
Web fonts causing FOUT. Web fonts that load after the browser has already rendered text cause layout shifts when the font swaps in. Use font-display: optional or font-display: swap with proper fallback sizing to reduce the shift.
Dynamic content injected above existing content. Banners, cookie notices, and notification bars that appear above existing content cause layout shifts. Reserve space for these elements, or animate them in from outside the viewport rather than pushing content down.
The Priority Order
If you're starting from scratch:
- Fix CLS first — often quick wins with image dimensions and ad space reservations
- Address LCP next — image optimization and caching have the highest impact
- Tackle INP last — usually requires the most technical investigation
This order is based on effort-to-impact ratio. CLS fixes are often simple HTML/CSS changes. LCP improvements are mostly configuration and asset optimization. INP requires JavaScript profiling and optimization, which takes more time.
Frequently Asked Questions
Do Core Web Vitals affect my rankings a lot?
They're confirmed ranking factors but not dominant ones. Content relevance and backlink authority still outweigh page experience signals for most queries. However, when two pages are otherwise similar in quality, CWV is a tiebreaker — and failing all three metrics creates a meaningful disadvantage.
My PageSpeed lab score is great but field data shows Poor. Why?
Lab data uses a controlled simulated test. Field data reflects real user conditions — varying devices, network speeds, and geographic locations. A user on a 3G connection in a rural area will have a very different experience than the lab's simulated fast connection. Fix the field data issues; they reflect actual users.
How often should I check Core Web Vitals?
After major site changes (new theme, significant JavaScript additions, homepage redesign), check immediately. Otherwise, a monthly review of Search Console CWV data is sufficient for most sites.
Marcus Reed is Senior Editor & Digital Strategist at High5Expert. He specializes in technical SEO and web performance optimization across a wide range of site architectures.
Discussion
11 commentsThe priority order at the end — CLS first, then LCP, then INP — matches exactly what I discovered by trial and error. CLS was the easiest to fix (image dimensions and ad placeholder space) and I saw it go from 0.23 to 0.06 in one afternoon. LCP took more work. Haven't tackled INP yet.
Thank you! We do offer consulting and implementation services. Feel free to reach out through our contact page and we can discuss your specific needs.
INP replacing FID in March 2024 was a bigger deal than most people realized. FID only measured the first interaction; INP measures all interactions throughout the visit. I had a few pages that passed FID comfortably but fail INP because of a slow event handler on a button that appears mid-page.
Google's stance is that they care about content quality, not how it's produced. The key is to add genuine value, expertise, and unique insights. Use AI as a starting point, then add your own experience and perspective. Pure AI-generated content without human review is what gets flagged.
The web font CLS issue caught me out. My site looked fine visually, but CLS was 0.18 and I couldn't figure out why. Turned out to be a Google Font loading after initial render. Switched to font-display: swap with matching fallback metrics and CLS dropped to 0.04.
That means a lot coming from someone with your experience! Internal linking is often the most underutilized SEO tactic. Thanks for sharing!
Question: how much does Core Web Vitals actually move rankings versus content quality? I ask because I have a competitor who ranks above me with a PageSpeed score of 41, while mine is 89. Their content is older and thinner but they have more backlinks.
That's a great suggestion! We're exploring video content for our most popular guides. Stay tuned — it's on our roadmap.
You've described the current reality accurately: Core Web Vitals are a tiebreaker, not a primary signal. Your competitor's backlink authority is outweighing their poor performance. That said, if you were equal in backlinks, your CWV advantage would matter. The work isn't wasted — as your authority grows, the performance advantage becomes more valuable.
True
The explanation of the main thread blocking is the clearest I've seen. I had a developer tell me the site 'passed all the tests' but INP was 380ms. Turned out a third-party chat widget was running a timer on the main thread every 100ms. Removed it, INP dropped to 140ms immediately.
Yes, Google has officially confirmed they don't use the meta keywords tag for ranking. Focus your energy on title tags, meta descriptions, header tags, and most importantly — quality content that answers search intent.
A third-party timer running on the main thread every 100ms is a textbook INP killer. The 100ms interval means the timer is competing with every user interaction for main thread access. That's the kind of issue that's invisible until you actually profile it in DevTools — the chat widget looks harmless but it's running hidden background work constantly.
True
LCP was my biggest problem — hero image of 1.8MB in JPEG. Converted to WebP (340KB), added explicit width/height, moved to loading='eager'. LCP went from 5.1s to 1.9s. The biggest single optimization I've done on any site.
That's awesome to hear! Proper heading structure helps Google understand your content hierarchy, which is exactly what featured snippets need. Keep it up!
The field data vs lab data discrepancy question in the FAQ matches exactly what I'm dealing with. Lab scores are 91 but field data shows 'needs improvement' on LCP. Most of our traffic comes from mobile users on slower connections. Started serving smaller images specifically to mobile via srcset.
You should review them at least quarterly. Check your Google Search Console — if a page has good impressions but low click-through rate, that's a sign your meta description needs work. Also update them whenever you significantly change page content.
Can you run a site with excellent Core Web Vitals and still rank poorly? Asking because I've optimized everything and we're still on page 3 for our main keyword.
Fix critical technical issues first (broken links, missing meta tags, slow pages). A site with technical problems won't rank well no matter how good the content is. Once the foundation is solid, shift 80% of your effort to content creation.
Yes, absolutely. Core Web Vitals are one factor among many. Page 3 with excellent CWV suggests the primary gap is somewhere else — content authority, backlinks, or topic relevance. Run a content gap analysis against the pages ranking on page 1. Are they covering aspects of the topic you're not? Do they have significantly more referring domains? That's where the next opportunity is.
True