Core Web Vitals in 2026: A Practical Guide for Marketing Teams

Core Web Vitals measure three things: how fast your largest piece of content loads (LCP), how quickly your page responds to interaction (INP), and how much your layout jumps around while loading (CLS). Search engines use them as a ranking signal, but the bigger business impact is usually conversion rate, since slow, jumpy pages lose visitors before they ever see your offer.

The good news for marketing teams is that you don't need to read a waterfall chart to make progress. You need to know what the three numbers mean, where the usual problems come from, and which of them your own decisions cause. That's what this guide covers.

Why marketing teams should care, not just developers

It's tempting to treat Core Web Vitals as a purely technical checklist for the dev team. But most of the biggest wins come from content and marketing decisions: oversized hero images, autoplay videos, third-party marketing scripts (chat widgets, ad pixels, popups) that load before the page is interactive, and pop-ups that shift content around. Fixing these usually requires marketing and development working from the same priority list.

There's also a sequencing problem. Developers fix a slow page, then a new campaign launches with a heavy hero video, two extra tracking pixels and a newsletter pop-up, and the scores are back where they started within a month. Performance only sticks when the people adding things to the site understand what each addition costs.

The three metrics, in plain terms

• LCP (Largest Contentful Paint): how long until the main image or headline appears; aim under 2.5 seconds
• INP (Interaction to Next Paint): how long the page takes to respond after a click or tap; aim under 200 milliseconds
• CLS (Cumulative Layout Shift): how much visible content jumps around while loading; aim under 0.1

Each metric has three bands. LCP between 2.5 and 4 seconds "needs improvement" and anything over 4 seconds is poor. For INP the bands are 200 to 500 milliseconds and over 500; for CLS they're 0.1 to 0.25 and over 0.25. Google assesses them at the 75th percentile of real page loads, split by mobile and desktop. In plain terms: three out of four visits need to hit the target, not just your own visit on a fast office connection.

That last point catches a lot of teams out. Your site probably feels quick on a new laptop over fiber. Your customers are often on a mid-range phone with a patchy mobile connection, and that's the experience the field data reflects.

Fixing slow LCP: start with the images

Compressing and correctly sizing images accounts for the majority of LCP improvements we see. The largest element on most marketing pages is a hero image or a big headline, and the fixes are well understood:

• Serve images at the size they're displayed, in a modern format like WebP or AVIF, instead of a full-resolution export from the design tool
• Never lazy-load the hero image; mark it as high priority so the browser fetches it first
• Avoid hero sliders and carousels, which often hold back the main image until a script has loaded
• Check server response time, because slow hosting or an uncached page adds delay before anything else can start
• Watch A/B testing tools that hide the page until the experiment loads, since they push LCP back on every visit

Fixing INP and CLS: scripts, pop-ups and reserved space

INP problems are mostly about JavaScript. While the browser is busy running scripts, it can't respond to a tap, so the page feels stuck. On marketing sites the main culprit is usually third-party code: tag manager containers that have collected years of old tags, chat widgets, heatmap tools, consent banners and ad pixels. Auditing every third-party script on the page, and asking whether it actually needs to load immediately, tends to fix both LCP and INP. Remove the ones nobody uses, and delay the rest until the page has loaded or until someone interacts.

For CLS, reserving space for images, ads, and embeds before they load (instead of letting the page reflow once they appear) solves most of the problem. The other usual suspects are banners injected at the top of the page after load, cookie notices that push content down instead of overlaying it, and web fonts that swap in at a different size and reflow the text.

“A one-second improvement in load time is a conversion rate change you can usually see in your analytics within a week.”

Tools to check where you actually stand

• PageSpeed Insights: the fastest way to get real field and lab data for a single URL
• Chrome DevTools' Performance and Lighthouse panels: for diagnosing exactly what's slow, not just that something is
• Google Search Console's Core Web Vitals report: shows real user data across your whole site, not just one page
• WebPageTest: useful for comparing before/after results on a specific fix in detail

Field data (from real visitors, via Search Console or CrUX) and lab data (from a single test run) can disagree, and that's normal: lab data is best for debugging a specific page, while field data is what actually reflects your rankings and real user experience, so treat it as the source of truth when the two conflict.

One more thing to know about field data: PageSpeed Insights reports real-user data collected over a rolling 28-day window, and Search Console works from the same kind of data. After you ship a fix, the lab score changes immediately, but the field numbers improve gradually over the following weeks. Don't panic if Search Console hasn't moved a few days after a release.

A performance routine marketing and development can share

One-off fixes decay. The teams that keep good scores treat performance as part of how they publish, not a project they run once a year. That doesn't need to be heavy:

• Agree a simple rule for new third-party tools: someone owns it, there's a written reason for it, and it's reviewed after three months
• Set size limits for uploaded images and video, and have the CMS resize and convert images automatically where possible
• Test new landing pages on a real mid-range phone before a campaign goes live
• Check the Core Web Vitals report in Search Console monthly, and look into any group of URLs that slips from good to needs improvement
• Audit the tag manager every quarter and remove tags for campaigns that have ended

Where to start

None of this requires a full rebuild, though if the platform itself is the bottleneck, our Next.js vs WordPress comparison can help you decide. Most sites can meaningfully improve all three metrics through a focused performance pass of two to three weeks: image optimization, script auditing, and layout stabilization, all without touching the design. Start by running your five most important pages (usually the homepage, top landing pages and main service or product pages) through PageSpeed Insights, note which metric fails on each, and fix the page that brings in the most revenue first. Then put the routine above in place so the gains are still there six months from now. Our website maintenance checklist shows where these checks fit into a regular schedule, and our web development team can run the performance pass for you.

Frequently asked questions