Core Web Vitals are a set of three measurements that Google uses to describe how a web page feels to the people loading it: how quickly the main content appears, how fast the page reacts when someone taps or clicks, and how much the layout jumps while it loads. They exist because "page speed" had become a loose label for dozens of competing scores. Google's Chrome team picked a small number of user-centred metrics, attached fixed thresholds to them, and tied them to Search Console, PageSpeed Insights and, from 2021, to ranking in Google Search. For publishers and advertisers they now link search visibility, ad delivery and conversion.

Three metrics, three thresholds

The current set, as of September 2026, is Largest Contentful Paint (LCP), Interaction to Next Paint (INP) and Cumulative Layout Shift (CLS). Google's Search Central documentation assigns each a job: LCP covers loading, INP covers responsiveness and CLS covers visual stability.

LCP records the moment the largest image, video poster or text block in the viewport is painted. A good score is 2.5 seconds or less, and anything beyond 4 seconds is rated poor, with the band between labelled "needs improvement". Diagnostic tools split it into Time to First Byte (TTFB), resource load delay, load duration and render delay, as Cloudflare's Observatory launch showed. Across 9 billion events, Cloudflare found sites with a TTFB above 1,800 milliseconds were 70.1 percentage points less likely to achieve a good LCP than the average site.

INP measures the delay between a user interaction and the next frame the browser paints, covering input delay, event handler processing and presentation. According to Mozilla's developer documentation, every eligible click, tap and key press across the page's lifetime counts, scrolling does not, and pages with 50 or more interactions report roughly the 98th percentile to strip out outliers. Good is 200 milliseconds or less; poor starts above 500 milliseconds.

CLS is a unitless score built from the Layout Instability API. Each unexpected movement of visible content is weighted by how much of the viewport moved and how far. A score of 0.1 or less is good, above 0.25 is poor.

How a page passes

Thresholds are applied to field data rather than laboratory tests. According to web.dev, a page or site is classified as good on a metric when at least 75% of page views meet the good threshold, and tools should treat a page as passing only when all three metrics clear that bar at the 75th percentile. Google chose the percentile and the numbers together: its threshold research found that 1.5 and 2 second LCP targets were not consistently achievable at that percentile even for top-performing sites, while 2.5 seconds was.

The field source is the Chrome User Experience Report (CrUX). According to web.dev, CrUX draws only on Chrome users who opted in to share usage data. Its methodology page adds two eligibility tests: a page must be publicly discoverable, using the same indexability rules as search engines, and sufficiently popular, against a minimum visitor count Google does not disclose. Data is aggregated over a rolling 28-day window, which is why a fix deployed today takes weeks to show fully.

Search Console reports groups of similar URLs, while PageSpeed Insights shows CrUX figures above a separate Lighthouse laboratory run, and the two routinely disagree. A 2025 client case described by Noel Ceta showed a PageSpeed score rising from 65 to 92 while LCP worsened from 1.8 to 4.2 seconds, because lazy loading had been applied to the hero image, followed by a 20% traffic decline.

Where advertising enters the measurement

Ads are among the most common sources of layout shift, since they are requested asynchronously and inserted after surrounding content has rendered. According to the Google Publisher Tag (GPT) developer guide, reserving space for the largest creative size configured to serve is the most effective fix, at the cost of blank space when a smaller creative wins. Fluid slots, used for native formats, shift layouts by design. Google also built Publisher Ads Audits, a Lighthouse plugin whose cumulative ad shift metric isolates movement caused by ad iframes.

Loading and responsiveness are affected too. Client-side header bidding runs auctions in the browser, and the trade-offs between browser and server execution feed directly into LCP and INP. Google AdSense has made the link commercial: an email sent on March 24, 2026 pushed back/forward cache, the Speculation Rules API and AI-assisted debugging to publishers, citing a 27% rise in page views at Netzwelt and 9% mobile revenue growth at Yahoo! JAPAN News.

Origin and evolution

The Chromium team announced Web Vitals on May 5, 2020, in a post introducing the programme with LCP, First Input Delay (FID) and CLS as the core trio. Google's Ilya Grigorik told the W3C Web Performance Working Group ten days later that input measurement needed to expand beyond the first interaction. That limitation shaped the next four years.

Search integration followed. Google's April 2021 update set a mid-June start for the page experience ranking change, which combined Core Web Vitals with HTTPS, mobile-friendliness and the absence of intrusive interstitials. According to the Search Central blog, the mobile rollout began on June 15, 2021 and finished by the end of August 2021. Desktop followed from February 22, 2022, completing by the end of March 2022.

INP appeared as an experimental metric in May 2022. Google announced on May 10, 2023 that it would take over from FID, and INP formally replaced FID on March 12, 2024. Chrome then ended FID support across its tooling on September 10, 2024, with the CrUX BigQuery schema dropping the field from the 202409 dataset. Two months later Google removed the Page Experience report from Search Console on November 18, 2024, leaving the dedicated Core Web Vitals and HTTPS reports in place.

Why it matters for marketers

The clearest commercial evidence is correlational. Shopify's analysis, published on April 27, 2026 and written by Mateusz Krzeszowiak, found conversion about 3.5% lower for every additional 100 milliseconds of LCP across its merchant base, with 2.5-second stores converting roughly 30% below 1.5-second stores. The slowest 5% of stores were excluded. Those are platform-reported figures from aggregate buckets, not a controlled experiment.

Sector studies show the distance still to travel. A holiday-season analysis found 76% of retail and wholesale sites failing the assessment, while financial services passed at 62.5%. Across the wider web, the HTTP Archive's 2025 Web Almanac, using July 2025 CrUX data, put the share of origins with good Core Web Vitals at 48% on mobile and 56% on desktop, up from 32% and 41% in 2021.

Paid media inherits the same pages: a slow landing page loses visitors before an analytics tag fires, distorting conversion data and the bidding models trained on it. Behaviour analytics products now filter sessions by LCP, INP and CLS, placing the metrics next to engagement data.

Limitations and disputes

The ranking weight is small and Google says so. Martin Splitt described Core Web Vitals in a 2025 Search Off the Record episode as a tiny, minute factor in most cases. Google's page experience documentation states there is no single page experience signal and that aspects beyond Core Web Vitals do not directly lift rankings. A reading of Google's AI search guidance in May 2026 concluded that mid-range scores are not a primary ranking driver outside e-commerce.

The data source is narrow. CrUX reflects opted-in Chrome users only, so audiences heavy in Safari are represented through a different population than the one visiting. Cloudflare Radar data cited in 2024 put Chromium browsers at about 70% of visitor requests on its network, in the Speed Brain announcement. Low-traffic pages get no URL-level data, and the popularity threshold is secret.

Governance draws objections too. The thresholds are set by Google, which also operates the dominant search engine, the most widely used browser and a large share of publisher ad technology. The underlying browser APIs pass through web standards groups, but the pass-fail bands and the 75th percentile rule do not. Measurement blind spots have also persisted: single-page applications that change content without a full reload were, until 2026, largely measured on their first load only.

Not the same as

PageSpeed Insights score. The 0 to 100 number is a Lighthouse laboratory estimate from one simulated device. Core Web Vitals are field percentiles and can move in the opposite direction.

Page experience. Google's broader framing, which bundles Core Web Vitals with HTTPS, mobile usability, intrusive interstitial rules and ad placement. Core Web Vitals are one component.

Web Vitals. The umbrella programme includes supporting metrics such as TTFB and First Contentful Paint, which help diagnose problems but carry no pass-fail role in the core set.

Viewability. A Media Rating Council ad measurement standard about whether an impression could be seen. It shares none of the thresholds or data sources.

Recent developments

Browser coverage has widened. Safari 26.2, Firefox 146 and Chrome 143 shipped in December 2025, and Safari's addition of the LCP and INP interfaces completed cross-browser support for measuring the metrics, although CrUX itself remains Chrome-only. WebKit has kept reworking its engine since, rewriting its JavaScript module loader for the Safari 27 beta.

Chrome 151, rolling out from July 28, 2026, added soft-navigation and interaction-contentful-paint entries to the performance timeline, according to Chrome for Developers, so the metrics can be measured per route in single-page applications. Google's final origin trial notes said the use of that data in CrUX would be decided later. Industry baselines are hardening in parallel: Joost de Valk's website specification, published in May 2026, lists the three thresholds as required among 128 rules.

Timeline

  • May 5, 2020: Google's Chromium team announces Web Vitals, naming LCP, FID and CLS as Core Web Vitals.
  • May 2020: The Core Web Vitals report replaces the Speed report in Search Console.
  • April 2021: Google confirms a mid-June start for page experience ranking and launches the Page Experience report.
  • June 15, 2021: The page experience update begins rolling out in mobile search, completing by end of August 2021.
  • November 4, 2021: Google publishes the desktop timeline.
  • February 22, 2022: Desktop rollout begins, completing by end of March 2022.
  • May 2022: INP enters the Web Vitals programme as an experimental metric.
  • May 10, 2023: Google announces INP will replace FID in March 2024.
  • March 12, 2024: INP replaces FID as a Core Web Vital.
  • September 10, 2024: Chrome ends FID support in its performance tools.
  • November 18, 2024: Google removes the Page Experience report from Search Console.
  • December 2025: Safari 26.2, Firefox 146 and Chrome 143 complete cross-browser LCP and INP measurement support.
  • March 24, 2026: Google AdSense emails publishers performance guidance tied to Core Web Vitals.
  • April 27, 2026: Shopify publishes its merchant-wide conversion analysis.
  • May 2026: Joost de Valk publishes a 128-rule website specification with required thresholds.
  • July 28, 2026: Chrome 151 begins rolling out with soft navigation performance entries.

Summary

Who: Google's Chrome team defines the metrics and thresholds and publishes field data through CrUX, while Google Search applies them as part of page experience. Publishers, e-commerce operators, developers, ad operations teams and performance marketers are the parties measured.

What: Three field metrics covering loading (LCP, good at 2.5 seconds or less), responsiveness (INP, good at 200 milliseconds or less) and visual stability (CLS, good at 0.1 or less), assessed at the 75th percentile of opted-in Chrome page views over 28 days.

When: Announced in May 2020, applied to mobile ranking between June and August 2021 and to desktop in February and March 2022, with INP replacing FID on March 12, 2024 and soft navigation measurement arriving in Chrome 151 in July 2026.

Where: Measured in the browser through web performance APIs, aggregated in CrUX, and reported in Search Console, PageSpeed Insights, the CrUX API and BigQuery.

Why: Google wanted a small, stable set of user-centred signals in place of fragmented speed scores. For marketers, the metrics connect page build decisions, including how ads load, to search visibility, ad delivery and conversion, even though Google describes their ranking effect as minor in most cases.