Core Web Vitals in 2026: three readings, real thresholds
Core Web Vitals are three measurements of how a page feels to the people using it: how fast the main content appears, how quickly the page responds to a tap, and whether things jump around. Below: the thresholds as Google publishes them, what Chrome's 2026 field data shows, how much the scores count for ranking, and the fixes we'd try first.
- Good
- Needs improvement
- Poor
- 75th percentileThree in four real visits must hit the target
- Worst reading winsOne poor metric makes the whole group poor
- 28-day windowSearch Console reports the last four weeks of visits
- Two devicesMobile and desktop are measured separately
Thresholds from web.dev and Search Console Help, read 4 October 2026.
What Core Web Vitals measure
Core Web Vitals are three metrics Google uses to describe real visitors' experience of a page: Largest Contentful Paint (loading), Interaction to Next Paint (responsiveness) and Cumulative Layout Shift (visual stability). Google says they are used by its ranking systems, and that good scores don't guarantee a top ranking.
Core Web Vitals 2026 brought no change to the metrics themselves. Interaction to Next Paint replaced First Input Delay on 12 March 2024, and the thresholds on web.dev are the same as they were then. If you've read that Google tightened the loading target to two seconds this year, check the source: as of 4 October 2026, web.dev still says 2.5 seconds.
What did change is the data: Chrome's monthly pass rates show the share of sites with good responsiveness slipping through 2026.
Sources: web.dev, "Web Vitals" and "Interaction to Next Paint becomes a Core Web Vital on March 12"; Google Search Central, "Understanding page experience in Google Search results" (last updated 22 September 2026). All read 4 October 2026.
The observation chart: good, needs improvement, poor
Search Console sorts every reading into three bands, using these cut-offs from its Core Web Vitals report help page.
Three in four visits
Each reading is taken at the 75th percentile of page loads. If your LCP is 2.4 seconds there, at least three quarters of recorded visits loaded the main content that fast.
The worst reading decides
Search Console gives a group of pages the status of its poorest metric. Good INP and CLS can't rescue a poor LCP.
Phones and computers apart
Mobile and desktop are reported separately. A page can be good on desktop and need work on mobile, which is the common pattern for image-heavy sites.
Twenty-eight days of history
Search Console's figures cover the last 28 days of real visits, so a fix you ship today takes weeks to show fully in the report.
Sources: Search Console Help, "Core Web Vitals report"; web.dev metric pages for LCP, INP and CLS. Read 4 October 2026.
Read your own numbers against the chart
Run your page through PageSpeed Insights and look at the top section, which shows real-visitor data when Chrome has enough of it. Type the three 75th percentile figures below and the checker bands each one and gives the overall status the way Search Console would.
If PageSpeed Insights says there isn't enough real-user data, your page or site doesn't get enough Chrome visits to be measured on its own. Search Console handles that by grouping similar pages, or falling back to the whole site. Meanwhile, the lab score underneath is a diagnosis tool, not a verdict.
Where the web stands: Chrome's August 2026 numbers
Every month the Chrome UX Report publishes how many websites pass each reading, based on real Chrome users. The August 2026 dataset, released on 8 September 2026, covered 18,294,881 origins (roughly, websites).
Responsiveness has slipped all year
Good INP fell from 87.1% of origins in the January 2026 dataset to 85.3% in August. Chrome's release note for August calls the continued INP regression a cause for concern and says the team doesn't have a definitive reason. Loading is still the reading most sites fail, but INP is the one moving the wrong way.
Two cautions. They're for all websites Chrome measures, worldwide; Chrome doesn't publish an Australia-only pass rate in these notes. And a web-wide pass rate tells you nothing about your own pages; the checker above and your Search Console report do.
Source: Chrome for Developers, "CrUX release notes" (August 2026 dataset published 8 September 2026; January to July 2026 datasets for the trend), read 4 October 2026.
How much Core Web Vitals count for SEO
They count, but less than relevance. Google's page experience documentation (updated 22 September 2026) says Core Web Vitals are used by its ranking systems and recommends getting them to good. In the same breath it says good results in Search Console or third-party tools don't guarantee top rankings, and that chasing a perfect score purely for SEO may not be the best use of your time.
Google Search always seeks to show the most relevant content, even if the page experience is sub-par.
Google Search Central, "Understanding page experience in Google Search results", read 4 October 2026
Our reading of that: relevance comes first, and page experience helps when several pages are similarly useful. The same page says there's no single "page experience signal", and that other parts of page experience, such as HTTPS and avoiding intrusive pop-ups, don't directly lift rankings but make a site more satisfying to use.
So we'd fix Core Web Vitals for the people first. A slow hero image or a button that ignores a tap costs you enquiries whatever Google thinks of it. If a page has poor readings and thin content, fix the content before shaving milliseconds; if it's already the best answer on the page of results, the speed work may be what's left to do. Our technical SEO guide shows where speed sits among the other checks.
Field data versus lab data
The question we hear most is why Lighthouse shows a green score while Search Console reports poor pages. They're measuring different things.
Field data
Collected from real Chrome users on their own phones, laptops and connections, and reported in the Chrome UX Report. It's what Search Console shows and what the pass or fail status rests on.
- Where: Search Console, top of PageSpeed Insights
- Catch: needs enough visits to report
Lab data
One simulated load on a set device and throttled connection, run by Lighthouse. Repeatable, so you can change one thing and test again within minutes.
- Where: Lighthouse, Chrome DevTools, lower half of PageSpeed Insights
- Catch: can't measure real taps, so INP isn't in it
Lab tools can't measure Interaction to Next Paint, because nobody is tapping during a simulated load; Lighthouse reports Total Blocking Time as a stand-in for main-thread congestion. For real INP problems, field data tells you which pages are slow, and Chrome DevTools' Performance panel tells you why once you reproduce the tap yourself.
Sources: web.dev articles on Interaction to Next Paint and on improving INP (both last updated 2 September 2025), read 4 October 2026.
When the numbers you've been shown aren't the numbers that count
Three situations owners hit before anyone touches the code, each with the check that settles it.
Someone emailed us a red PageSpeed score of 38 and says the site is failing
CheckScroll to the top of the same PageSpeed Insights report. The part that says whether the Core Web Vitals assessment passed is field data from real Chrome users. The 0 to 100 number below it is a Lighthouse lab score from one simulated load on one device and connection.
ThenJudge the site on that field section and on Search Console's report. Google's PageSpeed Insights documentation calls a lab score of 90 or more good, but also says lab and field results can disagree in either direction. A red lab score beside a passing field assessment is tuning work, not an emergency. Ask the sender which field reading fails, on which group of URLs. If they can't say, the email was a sales hook.
The report lists dozens of opportunities and we can't tell which ones matter
CheckFind the reading that is actually amber or red in the field data. Then read only the lab diagnostics that point at that reading's element: the LCP image, the script holding up the main thread, the banner that shifts.
ThenLeave suggestions aimed at readings that already pass. Their estimated savings come from the same simulated load, so a big number beside a passing metric changes nothing for real visitors. Fix one cause, re-run the lab test to confirm it moved, and give the field data time to catch up.
We installed a speed plugin, the score jumped, and Search Console still says poor
CheckSearch Console covers the last 28 days of real visits, so it trails every change by weeks. The field data for a single URL in PageSpeed Insights is refreshed daily (still over a trailing 28 days), so it shows the trend a little sooner.
ThenTest by hand on a mid-range phone too. Plugins that delay or bundle scripts can lift the lab score while the menu, booking widget or chat bubble stays just as slow to answer a tap, and that tap is what INP records. If INP is the failing reading, remove plugins and widgets nobody uses before adding another plugin to manage them.
Sources: Google for Developers, "About PageSpeed Insights" (last updated 21 October 2024), read 5 October 2026; Search Console Help, "Core Web Vitals report", read 4 October 2026.
What to try first, reading by reading
These are the causes we find most often on WordPress and page-builder sites. Work on the reading that's actually poor in your field data, not the one a plugin dashboard nags about.
- Never lazy-load the hero image. web.dev's LCP guide is blunt about it: lazy-loading the largest image always delays it. Some themes and page builders lazy-load every image unless told otherwise.
- Tell the browser it matters. Add
fetchpriority="high"to the hero image, or preload it, so it starts downloading early. - Shrink the file. Serve a phone-sized WebP or AVIF to phones instead of a 4,000-pixel photo.
- Check server response time. web.dev notes that a slow Time to First Byte can make a 2.5-second LCP hard or impossible. Caching and better hosting fix this; image tweaks don't.
- Count the scripts. Every chat widget, slider, tracking tag and plugin adds JavaScript the browser must run before it can respond to a tap. Remove what nobody uses.
- Break up long tasks. web.dev's INP guidance is to split long JavaScript tasks and yield to the main thread so input can be handled between them. That's developer work, but worth asking for by name.
- Delay what can wait. Load non-essential third-party scripts after the page is usable.
- Test the real tap. Open the menu, the booking button and the quote form on a mid-range phone, because that's where INP is recorded.
- Give images and videos dimensions. Include width and height attributes, or reserve the space with CSS aspect-ratio, as web.dev recommends.
- Reserve room for late arrivals. Cookie banners, review widgets, maps and ad slots need a fixed space before they load.
- Watch web fonts. A late font swap can push text around; web.dev has a separate font best-practice guide for this.
- Keep the back button instant. Pages that qualify for the browser's back/forward cache restore without shifting, which web.dev lists as a highly effective CLS technique.
Sources: web.dev guides to improving LCP (updated 31 March 2025), INP (updated 2 September 2025) and CLS (updated 7 February 2025), read 4 October 2026.
A routine for a small business site
- Open the mobile report first. In Search Console, go to Core Web Vitals and open the mobile report. Note which reading is poor and how many URLs it affects.
- Test one example page. Click a URL group, pick an example URL and run it through PageSpeed Insights. The lab section names the elements involved, such as the LCP image or the shifting banner.
- Fix the template, not the page. Groups usually share a template, so one fix to the theme, header or page builder setting can clear dozens of URLs at once.
- Start tracking. Once the fix is live everywhere, use the report's Start Tracking button. Search Console watches the issue for 28 days and marks it fixed if it doesn't reappear.
- Re-check after plugin updates. A new slider, chat tool or theme update is the usual reason a passing site starts failing again.
If you'd rather someone else work through the list, our technical SEO service starts with the field data, and our WordPress SEO page covers the platform-specific parts.
Source: Search Console Help, "Core Web Vitals report", section on validating fixes, read 4 October 2026.
Core Web Vitals, answered
What are the Core Web Vitals in 2026?
They are Largest Contentful Paint (loading, good at 2.5 seconds or less), Interaction to Next Paint (responsiveness, good at 200 milliseconds or less) and Cumulative Layout Shift (stability, good at 0.1 or less). Each is measured at the 75th percentile of real visits, separately for mobile and desktop.
What does "Core Web Vitals assessment failed" mean?
In PageSpeed Insights it means at least one of the three field readings is outside the good band at the 75th percentile. The assessment passes only when all three are good. Look at which reading is amber or red and start there.
What is a good INP score?
200 milliseconds or less is good. Between 200 and 500 milliseconds needs improvement, and over 500 milliseconds is poor. INP looks at interactions across the whole visit, so a slow menu or form can fail it even if the page loads quickly.
Do Core Web Vitals affect SEO?
Yes, Google says they are used by its ranking systems. It also says relevance comes first and good scores don't guarantee top rankings, so treat them as one factor that helps when competing pages are similarly useful.
How do I check my Core Web Vitals?
Use the Core Web Vitals report in Google Search Console for your whole site, and PageSpeed Insights for a single URL. Both show field data from the Chrome UX Report when there's enough of it. Lighthouse in Chrome DevTools is for debugging.
How do I improve Core Web Vitals on WordPress?
Stop the hero image being lazy-loaded, compress and resize images, add caching, remove plugins you don't use and delay non-essential scripts. Then reserve space for banners and embeds. Test changes on a staging copy first, because speed plugins can break layouts.
Is there a Core Web Vitals WordPress plugin that fixes everything?
No single plugin fixes every cause. Caching and image plugins help with LCP, but slow INP usually comes from scripts other plugins added, and CLS often comes from theme layout choices. Measure first, then pick the tool for the reading that is failing.
Why does my PageSpeed score change every time I run it?
The PageSpeed Insights documentation says lab results vary from run to run even when nothing on the page changed, because of things like network availability, the test machine's hardware and other work competing for it. Run the test a few times and read the middle result, then judge the site on the field data, which covers 28 days of real visits.
Did Google change the Core Web Vitals thresholds in 2026?
No. As of 4 October 2026, web.dev and Search Console Help list the same thresholds as when INP replaced First Input Delay on 12 March 2024: 2.5 seconds, 200 milliseconds and 0.1.
Questions two to seven reflect Google's autocomplete suggestions for "core web vitals", "how to improve core web vitals", "core web vitals wordpress", "what is a good inp score" and "do core web vitals affect seo", checked 4 October 2026.
Want your readings checked?
The free audit reads your field data, names the poor reading and the template behind it.