Core Web Vitals
Three numbers Google uses to describe how a page feels to real people: how fast the main content shows up, how quickly it reacts to taps, and whether it jumps around while loading.
The three metrics
- Largest Contentful Paint (LCP), loading: when the biggest image or text block in the viewport finishes rendering. Good is 2.5 seconds or less.
- Interaction to Next Paint (INP), responsiveness: the slowest (roughly) tap, click or keypress response across the whole visit. Good is 200 ms or less. INP replaced First Input Delay on 12 March 2024.
- Cumulative Layout Shift (CLS), visual stability: how much visible content moves unexpectedly. Good is 0.1 or less.
Google also publishes two supporting metrics that aren’t “core” but explain the core ones: Time to First Byte and First Contentful Paint.
Why one slow visit in four is enough to fail
Each bar is one page view’s LCP. Google takes the 75th percentile: sort all visits, and read the value three-quarters of the way along. Increase the share of slow visits (older phones, bad networks) and watch it.
Median 1.8 s p75 2.2 s Good
Up to 25% slow visits, the median and p75 both stay good. Once more than a quarter of visits are slow, p75 jumps into the slow group and the page fails, even though most visitors were fine. That’s deliberate: the percentile makes you care about the slower tail of your audience.
How a page passes: the 75th percentile
Google doesn’t use your average. It takes the 75th percentile of all real page views over the last 28 days, split by mobile and desktop, from the Chrome UX Report. A page passes the assessment only when LCP, INP and CLS are all in the good range at p75. Put simply: at least three in four visits must be good on every metric.
If a URL doesn’t have enough traffic on its own, Search Console and PageSpeed Insights fall back to data for a group of similar pages or the whole origin.
Do Core Web Vitals affect rankings?
Yes, but modestly. Core Web Vitals are part of Google’s page experience signals. Google’s own Search Central documentation says good page experience doesn’t override relevance: a slow page with the best answer still outranks a fast page with a worse one. Where results are otherwise close, being fast is a tiebreaker, and it helps with everything that happens after the click: bounce rate, conversions and ad quality scores.
Two practical consequences: rankings use field data, so the Lighthouse score doesn’t matter directly; and the data is per-page (or page group), so fix the templates that carry your organic traffic first.
Where to see your numbers
- CoreVitals checker or PageSpeed Insights: field + lab for one URL, instantly.
- Search Console → Core Web Vitals report: every indexed URL grouped by template, with pass/fail counts. Needs site verification.
- CrUX dashboard / BigQuery / CrUX API: monthly and daily history for any origin with enough traffic.
- Your own RUM: the open-source
web-vitalsJavaScript library reports LCP, INP and CLS from your real visitors to any analytics endpoint.
A fixing order that works
- Look at field data first to decide which metric actually fails, on which device.
- Run a lab test to see why: the LCP element, long tasks, shifting elements.
- Fix the template, not the page: one change to a product template fixes thousands of URLs.
- Re-test in the lab to confirm, then wait about 28 days for field data to fully roll over. Track it in history.
Check it on your own page
Field and lab numbers for mobile and desktop, in about 30 seconds.
Questions people ask
What are the three Core Web Vitals?
Largest Contentful Paint (LCP) for loading, Interaction to Next Paint (INP) for responsiveness and Cumulative Layout Shift (CLS) for visual stability.
What are the Core Web Vitals thresholds?
Good: LCP at or under 2.5 s, INP at or under 200 ms, CLS at or under 0.1. Poor: LCP over 4 s, INP over 500 ms, CLS over 0.25. Everything in between needs improvement.
Is FID still a Core Web Vital?
No. Interaction to Next Paint replaced First Input Delay as the responsiveness Core Web Vital on 12 March 2024, and FID was removed from Chrome tools afterwards.
How long does it take for fixes to show up?
Lab tests show the change immediately. Field data uses a rolling 28-day window, so the improvement phases in over about four weeks.