What Is Largest Contentful Paint? LCP Without the Jargon
A while back I audited an ecommerce site whose developer swore the pages "loaded in 800 milliseconds." Technically true — the HTML arrived fast. The product photo, the thing customers actually came to see, showed up six seconds later. That gap between "something loaded" and "the thing you wanted loaded" is exactly what Largest Contentful Paint was invented to measure.
So, what is Largest Contentful Paint? LCP marks the moment the largest visible content element in the viewport — usually a hero image or a big block of text — finishes rendering. It's Google's proxy for "when did this page feel loaded to a human," and it's one of the three Core Web Vitals that feed into ranking.
What Is Largest Contentful Paint Measuring, Exactly?
As the page loads, the browser keeps track of the largest element painted so far — images, video poster frames, background images loaded via CSS, and block-level text. Every time a bigger element renders, the candidate updates. The final LCP is the render time of whichever element ended up largest once the page settled (measurement stops at first user interaction).
Common LCP elements, in rough order of frequency: hero images, featured product photos, large headings on text-heavy pages, and video poster images. On most sites I audit, it's an image about 80% of the time.
The Thresholds, and Why 2.5 Seconds
Google grades LCP at the 75th percentile of real user visits — meaning three out of four visits must beat the threshold, not your one lucky test run on office wifi.
| Rating | LCP | What it means in practice |
|---|---|---|
| Good | ≤ 2.5 s | Page feels responsive; no work needed here |
| Needs improvement | 2.5 – 4.0 s | Noticeable wait; worth fixing, not an emergency |
| Poor | > 4.0 s | Users are watching a blank space; fix this |
Is the ranking impact huge? No — content and links still dominate. But LCP correlates strongly with bounce and conversion, so a bad score costs you money regardless of what Google thinks of it.
The Four Things That Make LCP Slow
- Slow server response. If the HTML takes 1.8 seconds to arrive (TTFB), your LCP budget is blown before the browser renders a single pixel. Cheap shared hosting and unoptimized database queries live here.
- Render-blocking resources. CSS and synchronous JavaScript in the
<head>that the browser must download and parse before painting anything. - Slow resource load. A 2 MB hero image, an image served from a distant server with no CDN, or — my personal favorite audit find — a hero image with
loading="lazy"on it, which politely tells the browser to deprioritize the most important asset on the page. More on that trap in our piece on lazy loading. - Client-side rendering. If the LCP element only exists after a JavaScript framework boots up and fetches data, you've stacked every previous problem on top of each other.
Fixes Ranked by Payoff
Based on a few hundred audits, here's where the wins usually are, biggest first:
- Properly size and compress the LCP image. A hero image should rarely exceed 150–200 KB. Serve modern formats (WebP or AVIF), and you'll often cut a full second right there.
- Preload the LCP image with
<link rel="preload">and addfetchpriority="high". Remove any lazy-loading attribute from it. - Fix TTFB. Page caching on a CMS site is frequently a 500–1000 ms improvement for an afternoon of work. A CDN helps if your audience is far from your server.
- Trim render-blocking CSS and JS. Inline the critical CSS, defer the rest. This one's fiddlier and usually earns less than people expect, which is why it's third, not first.
Finding Your LCP Element (Two Minutes, No Guessing)
Before optimizing anything, confirm what the LCP element actually is — people routinely spend a week optimizing the wrong asset. PageSpeed Insights names it outright: run the URL, expand the LCP diagnostic, and it shows the exact element with a thumbnail. In Chrome DevTools, the Performance panel marks LCP on the timeline, and hovering the marker highlights the element on the page. Do this per template, not once per site: the homepage LCP might be a hero image while product pages paint a gallery photo and article pages paint the headline text. Each needs a different fix, which is why "improve LCP" as a single to-do item never gets done.
Lab Numbers vs. Field Numbers
PageSpeed Insights shows two datasets and people constantly quote the wrong one. The lab number comes from a simulated device on a throttled connection — useful for debugging, irrelevant for ranking. The field number comes from the Chrome User Experience Report (CrUX), real visitors over the past 28 days — that's what Core Web Vitals assessment actually uses. It also means any fix takes up to a month to fully reflect in the field data, so don't panic-refresh three days after deploying.
One page passing doesn't mean the site passes, either. Template-level problems — an unoptimized image component, a slow theme — drag down every page built on that template. That's why it pays to crawl the whole site and check performance page by page instead of testing the homepage and declaring victory. The homepage is usually the most optimized page you own; the interesting problems live three clicks deeper.
Frequently Asked Questions
What is a good LCP score?
2.5 seconds or less, measured at the 75th percentile of real user visits. Between 2.5 and 4 seconds counts as needing improvement, and anything over 4 seconds is rated poor. The percentile part matters: most of your visitors have to hit the target, not just your fastest ones.
Why is my LCP fine in testing tools but poor in Search Console?
Testing tools report lab data from a single simulated run, while Search Console reports field data from real Chrome users over 28 days. Real visitors have slower phones and worse connections than your test environment. Trust the field data and use lab runs only to diagnose what to fix.
Does fixing LCP improve rankings?
Modestly. Core Web Vitals act as a lightweight signal, more of a tiebreaker than a primary factor. The bigger payoff is behavioral: faster-rendering pages hold more visitors and convert better, which is worth the effort even if your rankings barely move.
Try WebsiteChecker.Tech Free
Run a free technical SEO audit on any website. Get a client-ready report in minutes.
Start Free Scan