Skip to main content

Measure LCP Sub-Parts To Improve Largest Contentful Paint

· 5 min read

The Largest Contentful Paint is the Core Web Vitals metric that websites most often fail.

To make optimizing LCP easier Google has started recommending looking at a breakdown of the metric called LCP sub-parts. By breaking LCP down into separate components you can more easily identify suitable optimizations.

What are the LCP sub-parts?

The Largest Contentful Paint metric measures how long after navigating to a page the biggest page element appears. The metric can be divided into four components:

  1. Time to First Byte: time to start loading the HTML document from the website server
  2. Resource Load Delay: time for the browser to discover the LCP image
  3. Resource Load Time: time for the browser to download the LCP image
  4. Render Delay: time for the browser to display the LCP image or other LCP element

If the LCP element is not an image then the Resource Load Delay and the Resource Load Time will be 0 milliseconds.

LCP sub-parts diagram

Time to First Byte (TTFB)

Time to First Byte measures how quickly a website server responds to the request for the HTML document. It includes time spent establishing a server connection, but not download time.

To reduce TTFB you can:

Resource Load Delay

In a lot of cases, the Largest Contentful Paint is triggered by an image element on the page. The Resource Load Delay measures how soon after receiving the server response the browser discovers the image resource and can start downloading it.

Resource Load Delay will be low if you just use a simple <img> tag in your HTML code. However, if you use libraries for lazy loading or if your website is a single-page JavaScript application then it may take longer before the browser finds the LCP image.

A request waterfall visualization can help you identify when a request is only made late on the page load process.

Request waterfall where the LCP image is discovered late

To improve LCP Resource Load Delay you can:

  • Reference the image directly in the page HTML
  • Use the fetchpriority="high" attribute to make sure it gets prioritized

Resource Load Time

If the LCP element is an image then the resource load time measures how long it takes to download this image. The large the image the longer it will take to download.

The size of the image is not the only factor impacting Resource Load Time. If the page is also loading other resources at the same time this can cause bandwidth competition on the network. Here you can see that a large MP4 video file is delaying the LCP image.

Request waterfall where the LCP image loads slowly

To improve LCP Resource Load Time you can:

Render Delay

The Render Delay LCP sub-part measures how long it takes for the LCP element to become visible. It applies to both LCP images and other types of LCP elements.

For LCP images, render delay may be caused when preloading the image resource. The image may be downloaded while there are still render-blocking resources being loaded or while content is hidden using CSS.

Request waterfall where LCP happens long after the LCP image is loaded

For other types of LCP elements Render Delay will generally be higher as no time is allocated to Load Delay and Load Time. Render-blocking resources and JavaScript-generated content will have an especially high impact here.

To improve LCP Render Delay you can:

How to measure LCP sub-parts with Lighthouse

To see the LCP breakdown in Lighthouse, find the Diagnostics section of the Performance score and look for the Largest Contentful paint Element audit.

This won't just show the LCP element but also the sub-part breakdown (although Lighthouse calls them phases).

Lighthouse screenshot showing LCP sub-parts

How to measure real user LCP sub-parts with DebugBear

DebugBear real-user monitoring (RUM) can report LCP sub-part timings for visitors on your website. The LCP metric dashboard provides a breakdown of the LCP metric.

You can also look at individual user experiences to see what LCP component slowed down their page load time.

DebugBear user experience with LCP details

DebugBear RUM provides detailed dashboards for each LCP sub-part.

DebugBear Render Delay dashboard

What LCP sub-part breakdown should you aim for?

If the LCP element is an image you want to keep the Resource Load Delay and Render Delay as short as possible. Google suggests that these two components together should not account for more than 20% of your overall LCP score.

That leaves 40% each for the TTFB and LCP Resource Load Time. Unlike the delay sub-parts these two components are almost unavoidable when an image is the largest content element.

What does the LCP breakdown look like for real users?

Google has provided LCP sub-part data for real users based on the Chrome User Experience Report (CrUX) dataset. You can find the data in this talk by Estela Franco.

This shows that for slow pages the Load Delay component, where the browser doesn't know about the LCP image yet, significantly contributes to the overall LCP.

For fast pages loading the image is rarely a significant LCP contributor.

Monitor your Core Web Vitals

DebugBear helps you test your website speed and monitor it over time. Start a free trial to find out how to speed up your website.

DebugBear LCP dashboard

Get a monthly email with page speed tips