EMPIREYO

← Knowledge base

Web development

Core Web Vitals in 2026: speed is not an extra

Core Web Vitals are three Google metrics: how quickly a page appears, how quickly it responds and whether it stays still. Google uses them in its ranking systems, but writes that relevant content comes first. This article explains what the thresholds mean, where to find your own figures and where the gains usually are.

 ·  by  ·  Reading time 13 min  ·  Web development  ·  updated 12 November 2026

In short

Core Web Vitals measure loading (LCP), responding (INP) and shifting (CLS). Google calls a page good when LCP stays within 2.5 seconds, INP up to 200 ms and CLS up to 0.1, measured with real visitors. Google uses them in ranking, but relevant content comes first. Most gains lie in images, loading order and less work for the browser.

The three measurements that count

Google’s own thresholds for a good experience (source: web.dev: Core Web Vitals). Hit all three, on mobile, in the field.

LCPwithin 2.5 s

Largest Contentful Paint: how fast the largest element is on screen. Feels like: the page is here.

INPunder 200 ms

Interaction to Next Paint: how fast the page responds to every click or tap. Feels like: it is listening.

CLSunder 0.1

Cumulative Layout Shift: how much the page jumps while loading. Feels like: it stays still where you click.

The three cards above show the thresholds Google calls good. In this article you will read what the three metrics measure exactly, why it is about real visitors, how much weight Google gives them, where the gains are per metric, what we measure ourselves and when you need to worry about them less.

What exactly do LCP, INP and CLS measure?

Each metric is about a different moment in a visit:

  • LCP, Largest Contentful Paint. How long it takes until the largest block of text or image is on screen. Usually that is a photo at the top or the main heading. For the visitor it feels like: the page is there.
  • INP, Interaction to Next Paint. How quickly the page visibly responds to a click, tap or key press. INP looks at the interactions during the whole visit and reports one of the slowest. In 2024 INP replaced the older metric FID, which only looked at the first interaction.
  • CLS, Cumulative Layout Shift. How much the content moves unexpectedly. Think of a button that jumps away because a bar is still loading above it.

Search Console puts pages into three groups per metric:

MetricGoodNeeds improvementPoor
LCPup to 2.5 sup to 4 smore than 4 s
INPup to 200 msup to 500 msmore than 500 ms
CLSup to 0.1up to 0.25more than 0.25

A group of pages gets the status of its worst metric. If LCP is good but INP is poor, the whole group is poor.

Why is it about real visitors?

The thresholds apply to the 75th percentile: out of every hundred page visits, at least seventy-five must stay within the threshold. That way the visitor with an older phone or a slow connection counts too, not just the average. Mobile and desktop are assessed separately; a page can score well on a computer and not on a phone.

The figures in the Core Web Vitals report in Search Console come from the Chrome User Experience Report (CrUX): anonymised measurements from real Chrome users over the past 28 days. That has a consequence. If a page has too few visitors, there is no value of its own for that page. Google then looks at a group of similar pages or at the whole site, and sometimes the report simply says there is no data.

Field data or a lab test: which counts?

There are two kinds of measurement. Field data comes from real visitors; that is what Search Console uses. A lab test is one you run yourself, with one device and one connection, for example in PageSpeed Insights or in Chrome's developer tools. PageSpeed Insights shows both: field data at the top if there is any, a lab test below.

Field dataLab test
Where it comes fromReal Chrome users, over 28 daysOne test, with one device and one connection
Where you see itSearch Console, and the field section in PageSpeed InsightsThe lab section in PageSpeed Insights, Chrome's developer tools
INPYesNo: without a real click there is no INP; Google names Total Blocking Time as a stand-in
What it is good forKnowing how visitors experience the pageFinding the cause and measuring an improvement straight away

Use them together. The field tells you whether there is a problem, the lab helps you find where it is. A lab score that varies a little each time is normal: network and device differ per test. Steer by the trend in the field, not by one lab test.

Which tool for which job? A simple split:

  • Search Console for the overview: which groups of pages are good, need improvement or are poor, on mobile and desktop.
  • PageSpeed Insights for one address: the field value of that page or of the whole site, and a lab test with the main causes.
  • Chrome's developer tools to click through a page yourself and see which task is holding the browser up.
  • Google's web-vitals library, if you want to collect the three metrics from your own visitors. Google calls this the easiest way to measure them all in JavaScript.

How much weight does speed carry in Google?

Google is fairly precise about this. Core Web Vitals are used by its ranking systems, and Google strongly recommends achieving good values. But there is no single page experience signal, Google writes: the systems look at many signals together. And Google always seeks to show the most relevant content, even if the page experience is below par. A good page experience can help when many pages answer the question about equally well.

Two of Google's statements are worth remembering. Good figures in the report do not mean you will reach the top. And chasing a perfect score just for findability may not be the best use of your time, according to Google. Meet the thresholds, and spend the rest of your time on what your visitors are looking for.

Speed is not only about Google either. A page that shifts or responds slowly is annoying for everyone who uses it, whether they come through Google or through a link in an email. How load time and the rest of the build fit together is covered in our article on page speed.

Where are the gains for LCP?

Google splits the time to LCP into four parts. Each part has its own fix:

PartWhat it isWhat helps
Time to first byteUntil the browser has received the first bit of the pageFewer redirects, caching, a server or CDN close to the visitor
Load delayTime between that first bit and the moment the large image starts loadingPutting the image in the HTML, with fetchpriority high, and never lazy-loading it
Load durationHow long downloading that image takesThe right size and a modern file format
Render delayTime between arriving and being on screenFewer stylesheets and scripts that hold back rendering

According to Google, almost all the time to LCP should be spent loading the page itself and that one large element. Every moment in which neither is loading is a gain you can take. Three causes are common: the largest image is lazy-loaded, it only appears after a script adds it, or a font holds back the text. Google recommends fetchpriority high on the likely LCP image, but on one or two images at most; beyond that, the hint stops working.

Styles and scripts can also hold up LCP. Google recommends inlining small stylesheets, not loading scripts without defer in the head of the page, and choosing a font display setting that does not block the text. If your site only builds its content in the browser, with JavaScript, LCP waits for that script. Google then names server-side rendering or prerendering as the fix.

How do you make a page respond faster?

According to Google, an interaction has three parts: the wait before the code starts, the time the code itself needs, and the time until the browser paints the new frame. What helps:

  • Break up long tasks. Let the browser do other things in between, so a next click does not have to wait.
  • Less work per click. In a click handler, do only what is needed to show the result; the rest can come afterwards.
  • A small page structure. A page with a great many elements makes every change more expensive to render.
  • Fewer scripts during loading. Scripts that run while the page loads can make an early click wait.

INP is the hardest to find in the lab, because it is about real clicks. Google recommends starting with field data that shows which interaction is slow. If you do not have that, click through the main routes yourself, including while the page is still loading.

How do you prevent layout shifts?

  • Give every image and video a width and height, or reserve the space with CSS. According to Google, images without dimensions are one of the most common causes.
  • Reserve space for what comes later, such as embedded video, maps, ads and notification bars. A cookie notice that lies over the page pushes nothing aside.
  • Watch your fonts. A web font that arrives late can make text shift. A matching fallback font limits that.
  • Animate with transform, not with properties that change the layout of the page.
  • Make sure the back button works fast. A page that the browser restores from memory on back and forward does not shift again.

Shifts within half a second of a visitor's click do not count. A button to load more can simply add something. If loading takes longer than that half second, reserve the space in advance, with a placeholder where the content will appear.

How long before an improvement shows?

Longer than you would like. A lab test shows an improvement immediately, but the field data looks back 28 days. A new version of your site therefore has to be used by real visitors for a while before the report changes. If you start the check in Search Console with Start Tracking, Google monitors for 28 days whether the problem still occurs. That check does not ask Google to recrawl your pages; it only looks at visitors' measurements.

So expect about a month between the fix and a green status. In the meantime, use lab tests to see whether you are on the right track, and note what you changed and when. Then you will know later which change made the difference.

What do you do first?

  1. Open the Core Web Vitals report in Search Console and start with mobile.
  2. Pick the group of pages with the worst status and the most visitors.
  3. Test one example page in PageSpeed Insights and see which metric fails.
  4. Fix the biggest cause per metric, using the tables and lists above.
  5. Re-measure in the lab, then start the check in Search Console with Start Tracking. Google then monitors for 28 days whether the problem is gone.
  6. Keep watching. Every new plugin, chat button or tracking code can make the figures worse again.

What do we measure on our own projects?

On our case study pages we give our own measurement of the homepage for each site: a phone screen, a slow connection, an empty cache and the median of three runs. That is a lab test, not a Google field value. The conditions differ slightly per measurement; the case study page gives them exactly.

SiteBuilt inUntil the page is on screenMeasured on
BasestepAstro0.58 s24 October 2026
snelrepareren.nlNode.js0.74 s24 October 2026
TaalProAstro0.78 s22 October 2026
Atiq TandartsenAstro0.82 s3 November 2026
Ikiz AdvocatenkantoorAstro0.98 s22 October 2026
Sittard TaxiNuxt1.23 s22 October 2026
Taxi GrouAstro1.63 s22 October 2026
ArtsenPoolNext.js1.7 s24 October 2026

When this does not apply

If your site has few visitors, Search Console may not give any field data at all. Then you work with lab tests and with what you see yourself on an ordinary phone. That is fine; the thresholds remain a good yardstick.

Behind a login, such as a customer portal or a dashboard, ranking plays no part. Speed still counts there, but for the people who work with it every day. And a page that hardly anyone visits does not need to be tackled first: start where most visitors arrive.

What we do not know: exactly how much weight Google gives the Core Web Vitals. Google gives no figure. Nor do we give figures on extra enquiries from a faster site, because we have not measured them.

Frequently asked questions

Where can I see my own Core Web Vitals?

In Search Console, in the Core Web Vitals report, separately for mobile and desktop. Those figures come from real Chrome users over 28 days. For one page at a time you can use PageSpeed Insights; it shows the field data if there is any, and a lab test.

My score varies per test. Is that normal?

For a lab test, yes: network and device differ each time. That is why the field counts, with measurements from real visitors over 28 days. Steer by that trend, not by one test.

Do I need a perfect score?

No. Google recommends meeting the thresholds for good, but also writes that chasing a perfect score just for findability may not be the best use of your time. Relevant content comes first.

Why do I see no data in Search Console?

Probably there are too few visits with Chrome for a reliable value. Google then looks at a group of similar pages or at the whole site, and otherwise the report shows no data. A lab test then gives you something to go on.

Can a slow site become fast without rebuilding everything?

Often for a large part, yes: images, loading order and unnecessary scripts deliver the first gains. If the slowness lies in the way the site is built, we will say so honestly.

Does your site meet the thresholds? Send us the address, and we will look at your field data and the biggest cause per metric. You can read more about how we build under web development. You work directly with Nawid, with a fixed team of five specialists.

Written with the help of AI, reviewed and edited by Nawid Rasa, who is responsible for the content. Sources are listed below the article. Spotted an error? Email hi@empireyo.com.

Get started

Does your site hit the thresholds?

Request the speed check. You get your field score per measurement, the cause and the shortest route to green.

Or message us directly

We usually reply within two working days. Prefer to call or email? That works too. How we handle your details is in our privacy statement.

Sources

Consulted on 5 November 2026.

Empireyo · Zwolle Take your phone, switch off the wifi and open your own homepage on mobile data. Count the seconds out loud. Everything over three you felt yourself too, right?