Core Web Vitals are three user-experience metrics: Largest Contentful Paint (LCP) for loading, Interaction to Next Paint (INP) for responsiveness, and Cumulative Layout Shift (CLS) for visual stability. They describe outcomes experienced by visitors; they are not a hosting-plan score.
Current metric definitions and thresholds
| Metric | What it measures | Good | Needs improvement | Poor |
|---|---|---|---|---|
| LCP | When the largest visible image or text block is rendered. | ≤ 2.5 seconds | > 2.5 to 4 seconds | > 4 seconds |
| INP | Responsiveness across a visitor's interactions, based on the delay until the next visual update. | ≤ 200 milliseconds | > 200 to 500 milliseconds | > 500 milliseconds |
| CLS | Unexpected visual movement of page content during a visit. | ≤ 0.1 | > 0.1 to 0.25 | > 0.25 |
Assess each metric at the 75th percentile: at least three quarters of measured page visits should meet the good threshold for that metric. Reports may separate mobile and desktop. These are Google/web.dev's recommended thresholds, not service-level promises or guarantees of a search position.
Field data and lab data answer different questions
- Field data summarizes real visits under their actual devices, networks, locations, page content, and interactions. PageSpeed Insights and Search Console use Chrome User Experience Report (CrUX) field data when it is available. CrUX reporting is aggregated and eligibility-limited; a URL may not have enough data to show a page-level result. Search Console groups similar pages, while PageSpeed Insights can show URL and origin views.
- Lab data comes from a controlled test, such as a Lighthouse run. It helps reproduce and diagnose issues, but a simulated visit is only one set of conditions. Lighthouse cannot measure INP without real user input; Total Blocking Time (TBT) is a lab diagnostic that can help investigate responsiveness, but it is not an INP result.
PageSpeed Insights can show both a CrUX field result and a Lighthouse lab result on the same screen. They can differ because the data sources, devices, networks, caching, page state, and user interactions differ. Compare like with like: check field data for mobile or desktop separately, note whether it is URL-level, origin-level, or a Search Console page group, then use a lab run to investigate a reproducible page state. If there is no field result, that means the report lacks eligible data; it does not prove the site is fast or slow.
A practical check
- Enter a public page in PageSpeed Insights and note the field-data availability, form factor, percentile, and date window shown in the report.
- Review Search Console's Core Web Vitals report for affected page groups and device type if the property is available to you.
- Run Lighthouse on a representative page state to inspect lab diagnostics. Test interactions and post-load behavior separately; a single initial load will not reproduce every visitor journey.
- After changes reach visitors, revisit field data over time. CrUX uses a rolling 28-day aggregation, so it does not instantly reflect a code change.
What these metrics do not tell you about hosting
A poor LCP can involve server response time, redirects, network conditions, render-blocking resources, images, fonts, or page structure. INP can be affected by JavaScript work and the interaction being tested; CLS can result from images without reserved dimensions, injected content, or fonts. A hosting provider is one possible factor among many, and these metrics alone do not isolate its contribution. Compare the same workload, configuration, device and network conditions, and enough real visits before drawing a provider conclusion. This guide does not rank hosts or promise that changing plans will improve a metric.
For related capacity assumptions, see the hosting needs estimator and the bandwidth limits guide. Those tools do not measure Core Web Vitals.
Sources checked November 3, 2026
- web.dev: Web Vitals — metric definitions, recommended thresholds, and field versus lab measurement.
- web.dev: Core Web Vitals workflows with Google tools — PageSpeed Insights, CrUX, 75th-percentile reporting, and tool limitations.
- web.dev: How the Core Web Vitals metrics thresholds were defined — threshold categories and percentile interpretation.
- Chrome UX Report: CrUX tools and data scopes — update windows and differences among CrUX, PageSpeed Insights, and Search Console.
Impact — TODO: apply/approval and configure referral URL — referral link not configured.
CJ Affiliate — TODO: apply/approval and configure referral URL — referral link not configured.
PartnerStack — TODO: apply/approval and configure referral URL — referral link not configured.