Skip to content
horyond.com Web, mobile & AI studio --:--:-- Paris
Book a call
All articles

Core Web Vitals: the technical foundation that helps (or holds back) SEO

Core Web Vitals measure what your visitors actually experience: loading time, responsiveness and page stability. Here is what the official thresholds say, how much they really weigh in Google and the technical choices that keep them good over time.

Core Web Vitals are three measures of your visitors' real experience: LCP (how long the main content takes to appear), INP (how quickly the page responds to clicks, taps and key presses) and CLS (how visually stable the page is). Google states that its ranking systems use them, but that it always seeks to show the most relevant content, even when the page experience is weaker. They help SEO without ever replacing substance, and a slow or unstable site mostly holds itself back: its visitors leave before reading.

This article covers the official thresholds, how to measure them without looking at the wrong number, and the technical choices that make the difference. Every value quoted comes from Google and web.dev pages, linked at each step.

What exactly do LCP, INP and CLS measure?

According to web.dev, the reference documentation from the Chrome team, LCP (Largest Contentful Paint) measures loading performance, INP (Interaction to Next Paint) measures interactivity and CLS (Cumulative Layout Shift) measures visual stability. LCP is the moment the largest visible element, often a header image or a headline block, finishes rendering.

INP assesses the page's overall responsiveness to interactions: a mouse click, a tap on a touchscreen, a press on a physical or onscreen keyboard. Hovering, zooming and scrolling are not counted. INP replaced FID (First Input Delay) and became a stable Core Web Vital metric in 2024, according to the same web.dev page.

CLS reflects unexpected layout shifts: the button that moves down just as you click, the text that jumps when an image arrives. Shifts that happen within 500 milliseconds of a user input are not counted, web.dev explains: opening a menu with a click is not penalised.

The official thresholds, metric by metric

  • LCP: good up to 2.5 seconds, needs improvement up to 4 seconds, poor beyond (Search Console Help).
  • INP: good up to 200 milliseconds, needs improvement up to 500 milliseconds, poor beyond (web.dev).
  • CLS: good up to 0.1, needs improvement up to 0.25, poor beyond (Search Console Help).
  • These thresholds are read at the 75th percentile of page loads, segmented across mobile and desktop (web.dev).

How much do Core Web Vitals really weigh in Google ranking?

They count, but they do not decide alone. Google Search Central's page on page experience states that Core Web Vitals are used by its ranking systems, while specifying that there is no single page experience signal, and that good results in the reports do not guarantee ranking at the top of the results.

The same page recalls that Google always seeks to show the most relevant content, even if the page experience is sub-par. On its dedicated Core Web Vitals page, Google highly recommends achieving good scores for success with Search. The sensible reading is this: weak content does not climb thanks to a fast LCP, but between two pages that answer the question equally well, experience is part of what Google seeks to reward.

The real brake is often somewhere other than ranking. A page that takes long to appear or moves under the finger loses its visitors before they read, fill in a form or place an order. That is why these measures matter as much for a custom website as for an online store.

Why doesn't your PageSpeed score tell the whole story?

Because there are two kinds of data, and Google only judges one of them. PageSpeed Insights shows field data, drawn from the Chrome User Experience Report (CrUX) and aggregated over the previous 28 days, alongside lab data, produced by Lighthouse in a simulated page load on a single device with fixed network conditions.

The Core Web Vitals assessment is based on field data: a page or origin passes when the 75th percentile values for INP, LCP and CLS are all good, or, when there is not enough INP data, when LCP and CLS are. The lab score remains very useful for diagnosing and fixing, but it may not capture the real bottlenecks your visitors face on their own phones and networks.

The Search Console Core Web Vitals report follows the same logic: it groups URLs that offer a similar experience, gives each group the slowest status among its metrics, separately for mobile and desktop, and does not show a group that lacks data for both LCP and CLS. A young or lightly visited site may therefore only appear at origin level, or not at all.

How do you improve LCP?

The LCP optimisation guide breaks this time into four parts: the server's time to first byte (TTFB), the delay before the browser starts loading the main resource, the duration of that load, and the element render delay. Each part has its own causes, and so its own fixes.

The recommended levers for LCP

  • Make the main resource discoverable in the initial HTML, rather than found late by a script or stylesheet (web.dev).
  • Give high priority (the fetchpriority attribute) to the image likely to be the LCP element, without applying it to more than one or two images.
  • Never lazy-load the LCP image: it always adds an unnecessary delay.
  • Serve resources from a content delivery network (CDN) close to visitors, ideally from the same origin as the HTML.
  • Reduce render-blocking resources: trim or inline critical stylesheets, defer non-essential scripts.

How do you make a page more responsive (INP)?

According to web.dev, an interaction breaks down into three parts: the input delay before handling starts, the processing duration of the code it triggers, and the presentation delay until the next frame is shown. Each of these grows when the browser's main thread is already busy, for example with JavaScript running in long tasks.

The INP optimisation guide recommends breaking up the work in event callbacks into separate tasks, limiting what runs before the next frame to the visual update it needs, and keeping the DOM to a reasonable size, since a large DOM makes every update expensive. In practice, this is decided when choosing the building blocks of the site: every added script, widget and tracking module weighs on responsiveness.

How do you stop the page from jumping (CLS)?

The CLS optimisation guide lists very concrete causes, and most of them are best handled when the site is built rather than afterwards.

The most common causes of layout shifts

  • Images and videos without dimensions: web.dev recommends always setting their width and height attributes, or reserving space with the CSS aspect-ratio property (web.dev).
  • Ads, embeds and late-loaded blocks: reserve space for them (minimum height or ratio), without collapsing it when nothing loads.
  • Web fonts: plan suitable fallback fonts, preload critical fonts and consider font-display optional.
  • Animations on the top or left properties: prefer transform, which does not trigger re-layout.
  • Going back: making pages eligible for the back/forward cache (bfcache) avoids replaying shifts when users return to a page.

Does a fast site stay fast on its own?

No. Core Web Vitals are measured on real visitors, over a rolling 28-day window according to PageSpeed Insights: a heavier header image, an added tracking script, a banner that pushes content down, and the score slips without anyone deciding it. Performance is built at the start, then monitored.

It is one of the arguments for custom work over a template, which we compare in custom website or template: you choose each building block, and so each performance cost. It is also why every Horyond Agency service includes a support subscription: hosting, client dashboard, support, security and improvements. A change to the site is then planned with its effect on experience in mind, instead of being discovered in Search Console weeks later.

Price guidance is on the pricing page, with a website subscription from 130 EUR per month. To review the performance of your current site or a new project, the simplest step is a free discovery call by video.

Frequently asked questions

Do you need a score of 100 on PageSpeed Insights?

No. The score out of 100 comes from Lighthouse lab data. What Google assesses is field data: LCP, INP and CLS at the 75th percentile of your real visitors. A site can pass the Core Web Vitals assessment without showing 100 in the lab.

Does FID still count?

No. INP replaced FID and became a stable Core Web Vital metric in 2024, according to web.dev. The three current metrics are LCP, INP and CLS.

My site does not appear in the Search Console Core Web Vitals report: is that a problem?

Not necessarily. The report only shows a group of URLs if it has enough field data for LCP and CLS. A recent or lightly visited site may only appear at origin level, or not at all. PageSpeed Insights lab data can be used to diagnose in the meantime.

Are good Core Web Vitals enough to rank well?

No. Google specifies that good results do not guarantee ranking at the top and that it first seeks to show the most relevant content. Core Web Vitals are a foundation: they keep you from losing visitors, they do not replace content that answers the question.

Sources

Got a project in mind?

Let's talk it through on a free, no-commitment discovery call.