Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Glossary term

Core Web Vitals

Google's user experience metrics.
SEOconceptUpdated

Definition

The web.dev documentation defines Web Vitals as Google's initiative to provide unified guidance on the quality signals that matter to the experience of a page, and Core Web Vitals as the subset that applies to all pages and that every site owner should measure. The current set is Largest Contentful Paint (LCP) for loading, Interaction to Next Paint (INP) for responsiveness, and Cumulative Layout Shift (CLS) for visual stability. INP replaced First Input Delay as the responsiveness metric, which is why older material lists a different acronym. Each metric has a documented "good", "needs improvement" and "poor" range, and the documentation states that the set evolves over time as measurement improves, so the definitions are versioned rather than permanent, and a team should know which version its tooling reports.

How it works

The metrics are collected in the field, meaning from the browsers of actual visitors who have opted into Chrome's usage reporting, and aggregated in the Chrome User Experience Report. web.dev documents that a page or origin is assessed at a high percentile of page loads, so a site passes only when the large majority of visits meet the threshold, not the average visit. Lab tools such as Lighthouse simulate a load on a fixed device and connection, which is useful for diagnosis but is not the assessment, and INP in particular cannot be measured without real interactions. Field and lab numbers therefore disagree routinely, and only the field numbers count.

MetricWhat it measuresTypical causes of a poor score
Largest Contentful PaintTime until the largest visible content element has renderedSlow server response, render-blocking resources, large or late-loaded hero images
Interaction to Next PaintDelay between a user interaction and the next visual update, across the visitLong JavaScript tasks, heavy third-party scripts, expensive event handlers
Cumulative Layout ShiftHow much visible content moves unexpectedly during the visitImages and embeds without reserved space, late-injected banners, web fonts swapping

Search Console groups URLs by status in its Core Web Vitals report, mobile and desktop separately, which is where most site owners first meet the assessment. The Core Web Vitals guide covers diagnosing each metric and the template-level fixes that move many URLs at once; the technical SEO hub places the metrics among the other crawl, index and rendering topics.

In practice

The assessment is per URL group and per device, so a site can pass on desktop and fail on mobile, or pass its articles and fail its product pages. The practical reading is: find the template whose group fails, identify which of the three metrics is failing, and fix the cause in the template rather than page by page. Because the thresholds are fixed and documented, a team can set a target without inventing one. The speed and mobile landing page guide covers the conversion side: a page that loads slowly or shifts under a visitor's thumb also loses conversions (CVR), whatever Search does with the score.

Illustrative example — not a real result

Reading an assessment for a product template

A retailer's product pages are grouped as "poor" on mobile. The field data shows LCP at the 75th percentile of 3.9 seconds against the documented "good" threshold of 2.5 seconds, INP at 180 milliseconds against a threshold of 200, and CLS at 0.04 against 0.1. Only loading is failing. The largest element on every product page is the hero image, served at full resolution and loaded after a chain of scripts. The team adds responsive image sizes, preloads the hero and defers the scripts; the next report shows LCP at 2.3 seconds and the group moves to "good", with the other two metrics unchanged because nothing about them was touched.

What it is not

Core Web Vitals are not a ranking score and not a substitute for relevance: Google describes page experience as one signal among many, and a slow page with the best answer still outranks a fast page with a poor one. They are not a Lighthouse performance score, which is a lab composite with its own weighting. They are not the whole of web performance; metrics such as Time to First Byte and First Contentful Paint are documented as supporting diagnostics, not Core Web Vitals. And passing is not a one-off: the field data is a rolling window, so a regression in a release shows up weeks later as a changed status, long after the release that caused it has been forgotten.

Search engine optimisation (SEO) is the discipline that treats the assessment as a page-experience signal, and conversion rate (CVR) is the business metric that moves when loading, responsiveness and stability improve, which is why the same work is owned by both the SEO and the conversion teams. Technical terms such as the Chrome User Experience Report, field versus lab data and the Search Console report are explained in the technical guide linked above. Crawl budget and indexing, the other half of technical SEO, are independent of page experience: a page can be crawled and indexed perfectly and still fail the assessment, and a page that passes can still be left out of the index for reasons that have nothing to do with speed.

Common mistakes

  • Optimising to a lab score while the field assessment stays poor, because real devices and connections are slower than the simulation.
  • Fixing pages one at a time instead of the template that produces the whole failing group, which is slow and leaves the next page with the same fault.
  • Reading the average when the assessment is made at a high percentile of visits, so a fast median hides a slow tail.
  • Treating a pass as permanent and not checking the report after each release that touches scripts, images or layout, so a regression is found by the status change rather than by the team.
  • Ignoring INP because an older responsiveness metric was passing; responsiveness is now measured across the whole visit rather than on the first input.

More terms to look up

Where this term is used

Sources

  1. Web Vitals — web.dev

Related