Largest Contentful Paint (LCP) measures how long the biggest image or text block in the first screen takes to appear. Google rates 2.5 seconds or less at the 75th percentile as good and over 4 seconds as poor. On most slow sites the delay comes from the server and late image discovery, with image size a minor factor.
That last point surprises most teams. The standard advice for years has been to compress the hero image and move on, and plenty of B2B sites have done exactly that and still fail. Google's own Chrome field data shows where the time goes, and Google added a new wrinkle on October 13 with experimental metrics for single-page apps. Here is how the metric works, why it slips on phones, and what to fix first.
Largest Contentful Paint measures when the main content lands
Largest Contentful Paint marks the moment a visitor can see that the page is useful. It reports the render time of the largest image, text block or video visible in the viewport, counted from the moment the user started navigating, and that clock includes redirects, connection setup and server response time (web.dev). The target is 2.5 seconds or less for at least 75 percent of page loads, measured separately for mobile and desktop.
On most sites the LCP element is a picture. The HTTP Archive's 2025 Web Almanac, based on July 2025 HTTP Archive and Chrome UX Report data, found an image as the LCP element on 76 percent of mobile pages and 85.3 percent of desktop pages, with text making up nearly all of the rest (HTTP Archive Web Almanac). On a B2B homepage that usually means the hero image or the headline beside it. LCP covers the loading half of the experience. Its companion metric, Interaction to Next Paint, covers what happens once someone starts clicking.
A good LCP score is harder to hold on mobile
Phones are where Largest Contentful Paint fails. The same Almanac data shows 74 percent of desktop pages with a good LCP against 62 percent on mobile, and mobile pages are nearly twice as likely to be poor, at 13 percent versus 7 percent (HTTP Archive Web Almanac). LCP is also the weakest of the three Core Web Vitals on phones: 77 percent of mobile sites pass Interaction to Next Paint and 81 percent pass Cumulative Layout Shift.
The business stakes are easiest to see on lead-generation pages. Renault and its data partner fifty-five analyzed more than 10 million landing page visits across 33 countries and found that a one-second Largest Contentful Paint improvement was associated with a 13 percent increase in completed lead forms on pages loading in around a second (web.dev case study). That figure comes from a regression on 2020 and 2021 data, so read it as a strong correlation. A newer before-and-after example is ecommerce platform Nuvemshop, which raised its share of stores with good LCP from 57 to 96 percent and recorded an 8.9 percent rise in mobile conversion rate from Google organic visitors (web.dev case study). Speed is one of the levers we look at first in any B2B website conversion review.
LCP subparts show the image is rarely the problem
Every Largest Contentful Paint time splits into four subparts with no gaps: time to first byte, resource load delay, resource load duration and element render delay. Google's 2024 analysis of Chrome field data shows which one hurts. For the median origin with poor LCP (on pages whose LCP element is an image), time to first byte runs 2,270 milliseconds, the wait before the browser even requests the LCP image runs 1,290, the image download takes 350 and rendering takes 360 (web.dev). Google's analysis found that most origins with poor LCP spend less than 10 percent of their LCP time downloading the image.
Two conclusions follow. A slow server response alone can make a good score impossible, since 2,270 milliseconds leaves almost nothing of a 2.5 second budget, and where visitors are far from your servers, a CDN that serves pages closer to them is often the first fix. The second is that the browser frequently learns about the hero image far too late. Google calls resource load delay the overlooked culprit: the median poor-LCP site waits almost four times as long to start the image download as the download itself takes (web.dev). Google's own guidance for AI Overviews and AI Mode lists a great page experience among the SEO fundamentals that still apply (Google Search Central), so speed work also supports visibility in AI Mode.
The LCP image needs priority, and lazy loading takes it away
Late discovery of the Largest Contentful Paint image usually comes from a handful of build habits. About 16 to 17 percent of pages lazy-load their LCP image, a share that has not moved since 2024, which tells the browser to wait on the one image visitors are waiting for. Only 17 percent of mobile pages with an LCP image mark it with fetchpriority="high", and preload hints appear on just over 2 percent (HTTP Archive Web Almanac). Hero images set as CSS backgrounds add another step, because the browser cannot request them until the stylesheet loads (web.dev).
Nuvemshop's fix shows how simple the correction can be once the cause is known. Its team removed lazy loading from images at the top of the viewport, added fetchpriority="high" only to the likely LCP image, and dropped CSS transitions that delayed the first section from counting as visible (web.dev case study). The durable version of that fix lives in your components. If the hero component in your website design system ships with a real img tag, no lazy attribute and a priority hint, every new landing page inherits a fast LCP image. That is the kind of rule we build into technical SEO work so it survives the next redesign.
Soft navigations change Core Web Vitals for single-page apps
On October 13, Chrome began publishing experimental Core Web Vitals in its CrUX dataset that count soft navigations, the in-app page changes single-page apps make with JavaScript (web.dev). Until now, CrUX recorded only the first hard page load of an app session. Google says stable scores stay unchanged while the new versions are evaluated, and it aims to graduate them in May 2027, at which point Search Console will switch to the new data.
Most B2B marketing sites will see no difference in their Largest Contentful Paint numbers. About 70 percent of origins get fewer than 5 percent of page views from soft navigations, and only 7 to 11 percent use them for more than half. The change is large for app-style sites: among origins with 75 to 100 percent soft navigations on Android, the share with good LCP rises from 26 to 87 percent under the new measure (web.dev). Product and documentation sites built as apps should check CrUX Vis now. The new numbers do not excuse a slow first load, though. Client-side rendering that waits on all of its JavaScript turns into render delay, which is why server-side rendering or static HTML often has the LCP advantage (web.dev). The same architecture choice affects how AI crawlers read your links.
An LCP optimization plan for B2B marketing sites
Start LCP optimization with field data, then fix in the order the subparts suggest. Google highly recommends good Core Web Vitals "for success with Search" and says this aligns with what its core ranking systems seek to reward (Google Search Central). Pull the Search Console Core Web Vitals report and PageSpeed Insights for your top lead-generating templates, and note which element is the LCP on each. Then work down the list: cut server response time with caching and a CDN, put the hero image in the HTML as an img tag, remove lazy loading from it, add fetchpriority="high", and keep render-blocking CSS and scripts out of the way.
Then protect the gains, because Largest Contentful Paint drifts as pages change. A new hero video, a third-party script in the head or a CMS update can each add hundreds of milliseconds without anyone noticing. Set a mobile LCP budget per template, check field data monthly for the ten pages that bring in most leads, and add LCP to the checks in your website audit. If your demo or pricing pages sit above 2.5 seconds on mobile and you want a second opinion on where the time goes, talk to our team.
