Loading, responsiveness, stability: three measurements taken from your real visitors. What they mean, and how to get them into the green. Three metrics: LCP (loading, 2.5 s at most), INP (responsiveness, 200 ms at most) and CLS (visual stability, 0.1 at most). Google assesses them on data from real Chrome users, at the 75th percentile, separately on mobile and desktop. Since 12 March 2024, INP has replaced FID as the measure of responsiveness. A good score helps but doesn’t replace relevant content: speed opens the door, it doesn’t fill the room. Three measurements, three questions The Core Web Vitals are three metrics defined by Google to measure the real experience of a page. Each one answers a question visitors ask themselves without putting it into words. LCP, Largest Contentful Paint: “is it loading?” It’s the time it takes to display the largest visible element, often the main image or the headline. Good up to 2.5 seconds. INP, Interaction to Next Paint: “is it responding?” It’s the delay between an action (a click, a key press) and the visible update of the page. Good up to 200 milliseconds. CLS, Cumulative Layout Shift: “is it moving?” It’s the extent of unexpected layout shifts during the visit. Good up to 0.1. For a page to be rated good, these thresholds must be met for at least 75% of visits (the 75th percentile), measured separately on mobile and desktop. INP replaced FID in March 2024 Until 2024, responsiveness was measured with FID, First Input Delay, which only looked at the very first interaction, and only at the delay before it was processed. Since 12 March 2024, INP is what counts: it observes interactions throughout the visit, all the way to the display of their result. Websites that passed FID with ease found themselves in orange or red, especially those loading a lot of JavaScript. Field data or lab data: don’t mix them up The score out of 100 shown by Lighthouse is a lab measurement: a simulation, on a defined machine and connection. It’s valuable for diagnosis, but it isn’t what Google uses to assess your page. Google relies on field data from the Chrome User Experience Report (CrUX), collected from real Chrome users. You can check it in PageSpeed Insights, in the section devoted to real-user data, and in the Search Console Core Web Vitals report, which groups pages by issue. A website with little traffic may lack field data: you then have to rely on the lab and on your own measurements. The most common causes, and how to fix them A slow LCP almost always comes from a main image or video that’s too heavy, a server slow to respond, or stylesheets and fonts that block rendering. You compress images and serve them as WebP or AVIF at the right dimensions, preload the main image instead of lazy-loading it, and reduce whatever blocks rendering. A high INP points to a congested main thread: too much JavaScript running at the wrong time, third-party scripts (live chat, ad tracking, widgets) or long tasks, over 50 milliseconds, that delay the response to clicks. You break those tasks up, defer what isn’t essential and take stock of third-party scripts. A high CLS is visible to the naked eye: text that jumps while you read, a button that moves just as you click. The usual culprits are images and videos without declared dimensions, banners inserted after the fact and web fonts that change the size of the text as they load. You reserve space for each element before it loads. Fast doesn’t mean found The Core Web Vitals are among the page experience signals Google takes into account, but Google says it itself: it always seeks to show the most relevant content, even when the page experience could be better. Speed is a condition, not a strategy. Once all three metrics are green, the work continues: answering your customers’ questions precisely, structuring pages around their searches, earning links from recognised websites in your industry. A fast website opens the door; its content is what fills the room. Where to start Open the Core Web Vitals report in Search Console to see which groups of pages have issues, on mobile first. Test a representative page from each group in PageSpeed Insights to identify the LCP element and the heaviest scripts. Fix things starting with the pages that matter most to your business: the home page, the service pages, those receiving your ad campaigns. Check the effect in the field data, calculated over a rolling 28-day period: improvements don’t show up overnight.