LCP thresholds
web.dev sets these targets:
| Rating | LCP |
|---|---|
| Good | 2.5 seconds or less |
| Needs improvement | Over 2.5 and up to 4.0 seconds |
| Poor | Over 4.0 seconds |
Google measures this at the 75th percentile of page loads, split by mobile and desktop. So 3 out of 4 visits need to hit the target. LCP is one of three Core Web Vitals, with INP and CLS. web.dev says a page passes only if all three meet their targets.
What counts as the largest element
web.dev lists the element types LCP looks at:
- Images
- Images inside an SVG
- Videos
- Elements with a background image loaded through CSS
- Block-level elements that contain text
LCP only counts what is visible in the viewport when the page loads, so it changes between phone and desktop. On a phone, the largest element might be the headline. On a desktop, it might be a wide hero photo.
On a contractor landing page, the LCP element is often a big photo of a finished job or a crew, or the headline above the quote form.
How to fix a slow LCP
web.dev splits LCP into four parts, each with its own fixes:
- Time to first byte. Cut redirects and serve the page from a CDN close to your visitors.
- Resource load delay. Make the hero image findable in the first HTML, and give it high fetch priority. Never lazy-load it.
- Resource load duration. Size and compress the image, and use modern formats.
- Element render delay. Cut render-blocking CSS, avoid synchronous scripts in the head, and break up long tasks.
For most lead pages, the quick wins are the hero image and the scripts in the head. A heavy photo straight from a camera, or a slider that loads five images, can push LCP well past the target on a phone. So can a stack of tracking and chat scripts that load before the page draws.
One roofing company's ad page went from a ClickMend score of 42 to 93, and its load time from 4.1s to 1.8s.
A free ClickMend scan of your ad's landing page includes Core Web Vitals and names slow parts like unused JavaScript. See the landing page speed test guide for how to read the results.
Frequently asked questions
Should I lazy-load my hero image?
No. web.dev's LCP guide says, "Never lazy-load your LCP image." Lazy loading delays the image that LCP is measuring.
Does ClickMend measure LCP?
Yes, as part of Core Web Vitals, which make up 25% of the ClickMend score. It uses real-user CrUX field data when Google has it, and lab data if not.