Core Web Vitals and SEO: What They Measure, How Much They Matter

Core Web Vitals bring together three measurements Google uses to judge what your website is like for a real visitor: how long the main content takes to show up (LCP), how fast the page reacts when someone taps something (INP), and how much the layout jumps around unexpectedly (CLS). They do count for SEO, but as one signal among many; useful, relevant content still comes first. Where you really feel them is in the business, because people who get tired of waiting leave without calling.

The three metrics and Google's thresholds

Google publishes a “good” threshold for each one; past that line, the page needs work.

• LCP (Largest Contentful Paint). Measures how long it takes for the largest element in view to appear when the page opens: usually the hero photo, a background video, or the main headline. It's good at 2.5 s or less.
• INP (Interaction to Next Paint). Measures how long the screen takes to react when someone taps a button, opens the menu, or types into a form, and not just the first time: it watches interactions across the whole visit and reports one of the slowest. It's good at 200 ms or less. INP replaced FID in March 2024. FID only looked at the wait before the first interaction, so a site that passed under FID can fail under INP.
• CLS (Cumulative Layout Shift). Adds up unexpected shifts in the content: text that suddenly drops because a banner appeared above it, or the button that moves right as you go to tap it. It isn't measured in time but as a score, and it's good at 0.1 or less.

Keep in mind: Google doesn't grade the load you see from the office on a fast connection. It uses data from real visits, on mid-range phones and spotty mobile networks, and what counts is what most of those visits experience.

What breaks them on a business website

There's rarely a single culprit, just a pile of small decisions, each reasonable on its own, that slowly turn a quick site into a heavy one. Check these causes first:

• Heavy images. The hero photo uploaded straight from the camera or a stock site, several times larger than the screen needs. If it's also lazy-loaded even though it's the first thing people see, LCP shoots up.
• Plugins and page builders. Every add-on brings its own styles and scripts, and many load them on every page even when they're only used on one. The browser has to download and process all of it before it can show what matters.
• Third-party scripts. Live chat, tracking pixels, the embedded map, the reviews widget, the booking tool. Each one competes for the same phone processor, and when they stack up, the call button is slow to respond. That's INP.
• Web fonts. Several families and weights pulled from another server. Text shows up late or, worse, appears in a placeholder font and changes size when the real one arrives.
• Ads and pop-ups that move the page. Promo banners, cookie notices, sign-up pop-ups, or ad slots that get injected at the top after someone has started reading. Same with images, videos, and maps without set dimensions: they arrive late and shove everything down.
• A slow server. Overloaded hosting or no caching means the first response takes a while, and everything else starts late.

How to measure them without being technical

Two free Google tools are all it takes.

PageSpeed Insights

Paste in a page's address and check the mobile tab first. At the top you'll see field data, what real Chrome users experienced over the past few weeks, if your site gets enough traffic to show it. Below that is a lab diagnosis: a simulation that's very handy for tracking down causes, but it isn't what Google uses to evaluate you. Don't test only the homepage; check your main service page and your contact page too, since that's where people decide whether to reach out.

Search Console

The Core Web Vitals report groups your site's pages by status, separately for mobile and desktop, and tells you which metric is failing in each group. It's useful for spotting patterns: if every service page fails CLS, the problem is the template, not one stray page.

Two caveats. A site with little traffic may not collect enough field data, and then the report comes up empty; you go by the lab test, knowing it's an approximation. And because real-world data builds up over weeks, you won't see the change the day after a fix.

Where to start fixing them

• The main image, first. Size it to how it's actually displayed, use a modern format like WebP or AVIF, compress it, and have it load with priority instead of lazily.
• An inventory of plugins and scripts. What's installed, what's actually used, what can go, what should load only on the page that needs it, and what can wait until someone interacts, like the chat.
• Reserved space for everything. Width and height set on images, videos, maps, and ads, and nothing suddenly appearing above the text someone is in the middle of reading.
• Fewer fonts. A few families and weights, served from your own domain, with a fallback font of similar proportions so the swap goes unnoticed.
• Server and caching. Hosting that fits your traffic and cached pages so the first response arrives quickly.

Sometimes a site piles up so many patches that rebuilding the template makes more sense than stripping out more pieces. At Brito we decide with the data in front of us: sometimes a cleanup is enough, and sometimes it's worth investing in web design and development that's lightweight from the start.

How much they really count for SEO

Google confirms that its ranking systems take Core Web Vitals into account, as part of what it calls page experience. It's also clear that relevance and helpful content come first: passing all three thresholds won't lift you on its own, and missing them won't sink a page that's the most useful answer to a search.

Where they do tip the scales is when several pages answer equally well. And there's an effect no rankings report captures: the person who came from an ad or from the map and left before seeing your phone number. You lose that customer even if your position doesn't budge, which is why we treat speed as part of SEO and AI search and not as a separate test.

Be wary of anyone selling a perfect lab score as the goal. That score isn't the one Google uses, and chasing it at all costs can force you to cut features your business actually needs. The goal is to pass the three thresholds with real visitors and have the site do its job. If you suspect your site is holding back sales for other reasons, go over the signs your website is costing you customers.

Speed is just one piece. To see where all of your marketing stands, area by area and with what to fix first, answer a few quick questions. Take the free marketing scorecard.