Find out what your LCP element is
You cannot fix what you have not identified. In PageSpeed Insights, open the diagnostics for LCP and it names the element. In Chrome DevTools, the Performance panel marks it too. On a service business site it is often a hero photo, a slider image or a block of heading text.
Test on mobile, since the mobile layout can have a different LCP element than desktop. A background image set in CSS is especially common and harder for the browser to discover early.
Break the delay into four parts
LCP time is the sum of four stages: waiting for the server to answer, waiting for the browser to discover the resource, downloading it, and rendering it. Look at which stage is largest before changing anything.
If server response is slow, no image tweak will rescue you. If the image is discovered late, the fix is about priority. If it downloads for ages, shrink it.
Fix the server stage
Slow response often points to cheap hosting, no page caching, or heavy plugins building each page from scratch. Turn on page caching, remove unneeded plugins, and consider a better hosting plan. A content delivery network can also shorten the distance between your files and your visitors.
Fix the discovery and download stages
Use a real image tag for the hero rather than a CSS background where possible, so the browser finds it early. Preload the image or add fetchpriority="high". Never lazy load it. Serve a size appropriate to the screen with responsive image attributes, and use WebP or AVIF where your platform supports it.
Slider plugins are frequent offenders because they load several large images and scripts before showing the first slide. A single static hero often scores far better.
Fix the render stage
Large CSS and JavaScript files in the page head can block drawing. Remove unused styles, defer non-essential scripts, and inline a small amount of critical CSS if your tools support it. If your hero is text, make sure a custom font is not holding it back.
Re-test after each change and keep notes. Field data in the Search Console Core Web Vitals report takes weeks to update, so rely on lab tests in the meantime.
A worked example
Say a contractor's home page has an LCP of 4.8 seconds on mobile. The report shows the LCP element is a 1.9 MB hero photo set as a CSS background, and server response takes 1.2 seconds. The order of fixes: turn on page caching to cut server time, convert the hero to a properly sized WebP of about 150 KB in an image tag, and add high fetch priority.
Result: the browser finds the image at once and downloads a fraction of the data. A drop below the 2.5 second mark is realistic. Actual results will vary with your host, theme and visitors, so test again and compare.
Common questions
What is a good LCP score?
Google's published guidance treats 2.5 seconds or less as good, at the 75th percentile of real visits. Check the current thresholds in Google's documentation.
Why is my LCP different in each tool?
Lab tools simulate one device and network. Real-user data blends many devices and connections, so the numbers rarely match exactly.
Is LCP a ranking factor?
Page experience signals are part of how Google evaluates pages, but great content still matters more. Good LCP mainly helps visitors stay.
This guide is general information. Search engines change how they work and their guidelines; check Google's current documentation for your situation. Nobody can guarantee a ranking.