Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Core Web Vitals are Google’s field-based measures of loading speed, responsiveness, and visual stability. The current set is Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS). A page meets Google’s recommended targets only when, for both mobile and desktop separately, at least 75% of measured page loads are at or better than all three “good” thresholds: LCP 2.5 seconds, INP 200 milliseconds, and CLS 0.1.
These metrics describe real user experience and can inform Search performance, but a good result is not a ranking guarantee. Use field data to understand what visitors experience and lab tools to reproduce and diagnose problems.
What each Core Web Vital measures
| Metric | What it measures | Good | Needs improvement | Poor |
|---|---|---|---|---|
| Largest Contentful Paint (LCP) | Loading performance: when the largest visible image or text block is rendered. | ≤ 2,500 ms | >2,500–4,000 ms | >4,000 ms |
| Interaction to Next Paint (INP) | Responsiveness: how quickly the page produces visual feedback after user interactions. | ≤ 200 ms | >200–500 ms | >500 ms |
| Cumulative Layout Shift (CLS) | Visual stability: how much visible content moves unexpectedly during loading and use. | ≤ 0.1 | >0.1–0.25 | >0.25 |
The bands above are Google’s PageSpeed Insights classifications. They are diagnostic categories, not promises that a particular change will deliver a specific percentage improvement.
How Google evaluates a page
The 75th percentile matters
Google evaluates the distribution of observed page loads rather than a single average or the fastest visit. A page is in the “good” range for a metric when the 75th-percentile value meets that metric’s threshold. In practical terms, at least three quarters of measured experiences are at or better than the target.
#1 Best Overall
Mobile and desktop are separate populations
Review mobile and desktop independently. A page can pass on desktop while failing on mobile because of different devices, networks, processing capacity, and interaction patterns.
All three metrics count
Meeting two thresholds does not make the page pass. LCP, INP, and CLS must all be in the good range for the relevant device category.
Rank #2
Do Core Web Vitals affect Google Search?
Google Search Central highly recommends good Core Web Vitals for Search success and a good overall user experience. Google places them alongside other page-experience aspects and says these signals align with what its core ranking systems seek to reward. That is a qualified relationship, not a standalone ranking shortcut: good CWV scores do not guarantee higher rankings, and poor scores should be addressed alongside content quality, relevance, accessibility, security, and other Search considerations.
Free tools Windows power users keep installed
One-click scans. No signup required.
Field data versus lab data
Field data: what real visitors experienced
Field data is collected from real users under varied conditions. Device capability, network quality, other running processes, viewport, and the interactions a visitor performs can all change the result. Google uses real-user measurement to determine whether a site meets the recommended thresholds.
Lab data: a repeatable debugging environment
Lab tools run a controlled test that is useful while developing or comparing a change. They help you reproduce bottlenecks, inspect requests, and identify blocking work, but they do not replace field measurement.
Lighthouse cannot directly measure INP without user input. Its Total Blocking Time (TBT) is a lab proxy that can indicate potential responsiveness problems; TBT is not itself a Core Web Vital and should not be presented as observed INP.
Rank #4
Which tool should you use?
| Tool | Best use | Data scope | Important limitations |
|---|---|---|---|
| Search Console | Find site-wide and template-level CWV patterns after verifying ownership. | CrUX field data; groups similar URLs. | Ownership is required; grouped results may not describe every URL individually. |
| PageSpeed Insights | Inspect one page or an origin and see field and lab views together when available. | Page-level or origin-level CrUX plus Lighthouse lab data. | A page without enough samples may fall back to origin data; neither field nor lab data may be available when there is insufficient data. |
| Lighthouse and Chrome DevTools | Reproduce and diagnose development issues. | Lab measurements. | Controlled runs do not represent your entire audience; Lighthouse needs user input to measure INP. |
| Real-user monitoring (RUM) | Track detailed pageview, device, geography, or release segments. | Your own users’ field telemetry. | Requires implementation and responsible handling of collected data. |
CrUX availability depends on data sufficiency and eligibility. Before acting on a number, check whether it represents the individual URL, an origin, or a Search Console group. Compare field with field, lab with lab, page with page, origin with origin, and mobile with mobile.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
A practical measurement workflow
- Start in Search Console. Open the Core Web Vitals report for your verified property. Inspect the URL groups and identify templates or issue patterns affecting many pages.
- Open representative URLs in PageSpeed Insights. Record whether the field panel is page-level or origin-level, then review LCP, INP, and CLS separately for mobile and desktop. Use the Lighthouse section for reproducible clues.
- Reproduce the issue in Lighthouse or DevTools. Inspect the performance trace, network waterfall, main-thread tasks, rendering, and layout shifts. Treat TBT as a diagnostic proxy rather than an INP result.
- Segment with RUM when aggregate data is insufficient. Break results down by template, browser, device class, connection, geography, and release so that a single origin number does not hide a failing audience segment.
- Change one bottleneck at a time and recheck. Validate the change in lab runs, then wait for enough field observations to determine whether real-user percentiles improved.
How to improve LCP
LCP is often constrained by the path from the server to the browser and by work that delays rendering of the largest element. Google’s workflow guidance points to time to first byte (TTFB) and first contentful paint (FCP) as useful milestones.
Best Value
- Used Book in Good Condition
- Check server response time. A slow TTFB can delay every later loading milestone. Investigate application work, database waits, cache behavior, and hosting capacity before changing front-end code.
- Check delivery distance and routing. A content delivery network can help when geographic distance or origin delivery is the measured bottleneck. It is not an automatic fix for every LCP problem.
- Inspect redirects. Unnecessary redirect chains add round trips before the document and its largest element can load.
- Find render-blocking resources. CSS, scripts, fonts, or other resources that block parsing or painting may postpone LCP. Confirm the blocking relationship in a trace before removing or deferring anything.
- Test slower networks and devices. A desktop run on a fast connection can conceal the experience that drives the mobile 75th percentile.
How to improve INP
INP reflects the delay between a user interaction and the next visual update. Investigate the interactions that create long main-thread tasks rather than relying on a single synthetic score.
- Use DevTools traces to locate long JavaScript tasks during clicks, taps, typing, and other common interactions.
- Reduce unnecessary JavaScript, split expensive work, and avoid forcing large synchronous updates immediately after an input.
- Measure the resulting behavior with field data, because lab TBT can reveal blocking work but cannot substitute for observed INP.
How to improve CLS
CLS rises when visible content moves unexpectedly. Diagnose the element and timing of each shift in a trace or field report.
- Reserve space for images, video, advertisements, embeds, and other late-loading content so the layout does not have to reflow.
- Check fonts and injected components for late changes in size or position.
- Distinguish expected movement caused by a user action from unexpected movement during loading; the metric is intended to capture visual instability that visitors did not request.
Common interpretation mistakes
- Using one Lighthouse run as the site score. It is a lab sample, not a percentile of your audience.
- Calling TBT “INP.” TBT is a lab proxy; INP requires interaction observations.
- Applying origin data to every URL. PSI may show origin-level CrUX when a page lacks sufficient samples.
- Mixing device populations. Compare mobile to mobile and desktop to desktop.
- Buying a hosting or CDN service before diagnosis. Consider delivery changes when TTFB, network distance, or routing evidence shows that delivery is the limiting factor.
What a successful CWV program looks like
Use Search Console or RUM to find where real users struggle, PageSpeed Insights to establish the field-versus-lab context for representative pages, and Lighthouse or DevTools to isolate a cause. Prioritize the largest affected template and the metric with the clearest bottleneck, then verify both the technical change and the later 75th-percentile field result. Recheck definitions and tool availability before republishing guidance because Google’s metric set and reporting interfaces can change.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
