An Elementor site, rebuilt as a block theme
An Elementor site I rebuilt as a block theme went from 8.25 s to 2.03 s for its main content on a phone. Most slow WordPress sites need far less than a rebuild: the fixes below cover the photos, fonts and scripts I find on nearly every one.
If you have ever tested your WordPress site in PageSpeed Insights, you have seen the terms: LCP, INP, CLS, render-blocking resources, unused JavaScript. It is a lot. But for most sites the goal isn’t a perfect score. It is a site that feels fast, stable and easy to use for real visitors. That is what Core Web Vitals are trying to measure.
The site at the top had already been tuned as far as Elementor allowed. I rebuilt it as a block theme with the same design and content, and its main content appeared 6.22 s faster on a phone.
How this post was made. The figures come from my own client work and are checked against my notes. I drafted the post with help from AI, then edited it.
LCP, INP and CLS: the three Core Web Vitals
Google judges each one on real visits: a page passes when at least 3 in 4 visits are in the good range, so a fast result on your own laptop doesn’t mean much on its own.
Google’s ranges for each Core Web Vital
LCPMain content appears
INPPage reacts to a tap
CLSPage holds still
A page passes when at least 3 in 4 real visits are in the good range for all 3.
LCP: how long the main content takes to appear
Largest Contentful Paint is the moment the biggest thing on the first screen, usually the hero image or the headline, has loaded. Good is 2.5 s or less. Over 4 s is poor. It is often the most important of the three on a WordPress site, because it is how fast the page feels when someone opens it.
What I usually find behind a slow LCP:
- Large hero images and sliders
- Slow hosting or no page caching
- Heavy themes and page builders
- CSS that blocks the page from drawing
- Fonts that load too slowly
- The hero image being lazy-loaded, so the browser fetches it last
Images are the most common cause. On one small business site the header images alone came to 16.2 MB. Converting them to the right format, with no resizing and no visible change, took the home page from 6,017 KB to 1,041 KB.
Page weight before and after converting the photos
Home page 83% lighter
Equipment page 93% lighter
About page 83% lighter
Page weight measured logged out at phone width. Photos saved as PNG converted to JPEG at quality 90, checked pixel by pixel against the originals.
INP: how quickly the page responds
Interaction to Next Paint measures the delay between a tap or click and the page reacting to it. Good is 200 ms or less. Over 500 ms is poor. It can only be measured with real visitors, which is why PageSpeed’s lab test shows Total Blocking Time instead.
The cause is nearly always JavaScript:
- Page builder scripts
- Plugins loading their files on every page
- Tracking and tag manager scripts
- Chat widgets, popups and cookie banners
- WooCommerce scripts on pages that sell nothing
Tracking is the one people don’t expect. On a brand new site that scored 99 to 100 on mobile, adding the standard Google Tag Manager snippet dropped it to 75–85 and pushed the main content from 1.7 s to 4.5 s. Loading the tag manager after the visitor’s first tap or scroll brought all 13 pages back to 98–100.
Main content time on a phone, with Google Tag Manager
2.8 s faster
PageSpeed Insights, mobile. If you run Google Ads, visitors arriving from an ad still need the tag straight away, so they are exempt; otherwise the delay costs you the click data.
CLS: whether the page holds still
adds up how much things jump around while the page loads. Good is 0.1 or less. Over 0.25 is poor. It is why you sometimes tap the wrong link while a page is still loading.
- Images without set dimensions
- Web fonts swapping in at a different size
- Cookie banners and announcement bars pushing the page down
- Ads, embeds and form widgets that appear late
- Sliders and sticky headers
The fixes are usually small once you know which element moves. Each of these took about an hour or less:
Layout shift before and after 1 fix
Space kept for a spam-check box Desktop score 80 → 100
Space kept for images while they load Desktop score 70 → 92
Layout shift on desktop, 2 client sites, 1 fix each.
WordPress is rarely the problem
The problem is usually what gets added over time: a page builder, a plugin for every feature, analytics, forms, cookie tools, sliders, embedded videos. Each one seems small. Together they slow the site down, which is why performance tends to get worse gradually.
I tested this directly. On the same host and CDN, Astro beat my WordPress site by 520 ms, and my WordPress site beat a typical page builder site by 2.78 s. The full test is in Astro vs WordPress: A Speed Test on the Same Host.
Mobile main content time
What your PageSpeed score shows
A good score helps, but sites can score well and still feel slow, or score lower because of a few warnings while feeling fine. PageSpeed Insights shows two different things. The section at the top, when your site gets enough traffic, is real Chrome visitors over the last 28 days. The score underneath is one simulated phone on one connection.
The PageSpeed score
- Tests one simulated device and connection
- Flags technical warnings, useful and otherwise
- Can tempt you to chase 100
- Changes a little every time you run it
Real performance
- Comes from the people who visit
- Shows whether pages load, respond and hold still
- Points you at the improvements that matter
- Is what Google uses for Core Web Vitals
For a business site, a fast, clear, reliable experience matters more than a perfect score.
How to improve Core Web Vitals on WordPress
Not every PageSpeed warning deserves the same attention. This is the order I work in on most WordPress sites:
Test the pages that matter
The home page can be fine while service pages, posts or the shop are slow. Start with the pages that bring in leads and sales.
Find the actual LCP element
PageSpeed names it. Is it a hero image, a slider, a heading? Is it being lazy-loaded?
Size and compress images
The right size for the screen and a modern format. This is usually the biggest single saving.
Stop plugins loading everywhere
A form plugin only needs its files on the page with the form.
Delay what can wait
Tracking, chat and other third-party scripts can load after the first tap or scroll.
Give everything its space
Images, embeds, banners and fonts should take their space before they load, so nothing moves.
You don’t need to fix everything at once. Start with your most important pages and your mobile visitors.
Check your own site for free
My free health check runs these tests for you. It reads 8 pages of your site the way a visitor on a phone sees them, then sends a 2-page report in plain English covering 4 things:
- SpeedHow long your main content takes on real visitors’ phones, against Google’s “good” range.
- Google basicsPage titles, descriptions and how your pages look when shared.
- AI-readyWhether AI assistants and agents can read and use your pages.
- Trust and mobileSecure connection, and how easy you are to contact from a phone.
It ends with what is already working and what to fix first.

When to get help with WordPress speed
Some of this you can do yourself: removing plugins you no longer use, compressing large images, setting up caching, replacing a heavy slider. If the site still feels slow, or Search Console is reporting Core Web Vitals problems, it is worth a proper review. Especially if your site:
- Uses a page builder
- Has a lot of plugins
- Is slow on phones
- Runs WooCommerce
- Feels slow despite a decent PageSpeed score
- Has had optimisation work before without a clear result
- Brings in leads, bookings or sales
My performance audit finds what is holding the site back and turns it into a clear plan. If you would rather not think about it again, a maintenance plan keeps the site fast as it changes. You can see what the work did for other sites in the case studies.
FAQ
Core Web Vitals: common questions
Short answers to the questions I get asked most.
LCP of 2.5 s or less, INP of 200 ms or less and CLS of 0.1 or less, for at least 3 in 4 real visits. Pass all three and the page passes.
The test runs on a shared machine over a simulated connection, so it varies from run to run. Run it several times and look at the middle result, not the best or worst one.
They are part of how Google judges page experience, but relevant, useful content matters far more. A pass won’t lift a weak page above a better one. Speed matters most for the visitors you already have.
Google only reports real-visitor data once a page gets enough Chrome traffic. Until then, the lab test in PageSpeed Insights is the best guide you have.
Rarely. A well-built WordPress page passes comfortably. The slow part is usually the page builder and plugins, not WordPress. My Astro vs WordPress test has the numbers.




