New: remix a demo site in your browser – no sign-up, nothing to install. Try it

← Blog

Elementor Google Page Speed: A Diagnostic Guide

· 9 min read

Key takeaways

  • Read PageSpeed Insights metric by metric instead of trusting the overall score, since lab data on a simulated mobile device is the honest baseline.
  • Enable Elementor's Improved DOM Output, Improved CSS Loading, Optimized Image Loading, and Google Fonts Load Swap, then exclude the hero image from lazy loading to protect LCP.
  • Fix CLS by loading fonts locally with font-display: swap, wrapping embeds in fixed aspect-ratio containers, and avoiding animations that collapse or expand elements.
  • Audit add-ons with the Chrome DevTools Coverage tab and per-widget deletion tests, keeping only widgets used on live client pages.
  • Expect 70-90 mobile and 90+ desktop for a well-tuned Elementor site; if unused CSS and JS persist on every build, switching builders beats more tuning.

A slow Elementor site can be fixed, and the fix starts with reading PageSpeed Insights as a map rather than a grade. Each failing metric points to a specific Elementor setting, widget, or add-on, and this guide walks you through diagnosing and repairing each one, plus the cases where switching builders is the honest answer.

Why is my Elementor page speed slow according to Google PageSpeed Insights?

Elementor sites score poorly for structural reasons, not because you configured something wrong. The builder wraps every section in nested divs, loads its own CSS and JS on every page, and add-on widgets pile more of both on top. Google flags the result as a large DOM size and unused CSS and JavaScript, which are exactly the findings that dominate an elementor google page speed report.

What the builder itself adds before your content starts

Before your first headline renders, the page has already requested Elementor's frontend styles, its widget scripts, and anything your theme or add-ons inject. Each widget you drop in carries its own markup and often its own stylesheet. PageSpeed measures what ships, not what you intended, so a page that looks simple can still fail the metrics Google defines in its Core Web Vitals documentation.

Why scores differ between mobile and desktop

PageSpeed Insights runs a lab simulation of a mid-tier mobile device, so render-blocking requests and DOM depth cost far more than on your desktop test. The same URL can score 90 on desktop and 40 on mobile. That gap is the diagnostic signal, not noise. The same physics applies to any builder, including our own scroll-motion theme at https://midgroundwp.com/ — which is why we treat mobile lab scores as the honest baseline for client work.

How do I read PageSpeed Insights results for an Elementor site?

PageSpeed Insights gives you two different datasets. Lab data is a simulated test run on a Moto G Power with a slow connection. Field data comes from real Chrome users over 28 days. On a low-traffic client site, field data may be empty or thin, so trust lab data for diagnosing, then re-test after fixes.

Lab data vs field data: which to trust for a low-traffic client site

Ignore the overall score. Google uses it for marketing, not ranking. Diagnose metric by metric instead.

The three metrics that map directly to Elementor features

MetricWhat it measuresWhat Elementor influences
LCPTime until the largest element rendersImage sizing, lazy loading, hero widgets
CLSLayout shift during loadFont loading, ad-hoc spacing, animations
INPResponsiveness to clicksJS execution per widget

TTFB and FCP matter too, but hosting and theme control them more than the builder. The same discipline applies if you evaluate alternatives like Midground: judge metrics, not scores.

Which Elementor settings most affect page load speed?

Four settings move the needle: Improved DOM Output (Element > Settings > Features) flattens wrapper divs and shrinks the DOM size flag in PageSpeed Insights. Improved CSS Loading splits Elementor's stylesheet so pages load only the CSS their widgets use, cutting unused CSS. Optimized Image Loading applies lazy loading and proper sizing, and Google Fonts Load Swap sets font-display to swap so text renders without waiting on font files.

Lazy loading deserves its own caution: on images above the fold it can delay the LCP element rather than help it, so Elementor lets you exclude the hero image. Google's own guidance covers when lazy loading helps and when it hurts LCP.

Element > Features experiments worth enabling

Element Caching stores rendered page output server-side, so repeat visits skip widget rendering entirely. Turn it on once your layout is final.

Settings that make no measurable difference

Optimized Gutenberg Loading and the various asset "optimization" toggles rarely change scores on a simple page. If a toggle shows no PageSpeed delta after two runs, leave it alone and move on. Scroll-motion done without a builder is possible too; our Northern Lights demo shows the effect.

How do I fix LCP and CLS issues caused by Elementor?

LCP: the hero section checklist

Your largest contentful paint is almost always the hero image, and Elementor gives you three levers. First, upload a correctly sized WebP at the display width, not a 4000px original scaled down in CSS. Second, exclude the hero image from lazy loading: Elementor's lazy load plus a plugin's lazy load double-applies, and the browser delays the one image that matters most. Third, if you use a slider for the hero, don't. A single static image paints faster than any slider, and clients rarely miss the rotation.

Check the result by re-running PageSpeed Insights and looking at the LCP element named in the diagnostics. If it's still the hero after these fixes, the delay is server-side, not Elementor's markup.

CLS: fonts, embeds and entrance animations

Layout shift comes from things that arrive late and push content down. Load fonts locally with font-display: swap and a preload tag rather than pulling from Google's servers. Give YouTube and map embeds a fixed aspect-ratio wrapper so the iframe doesn't resize the page on load. And avoid entrance animations that start elements at zero height or off-screen; fade-and-translate in place, don't collapse and expand.

If this list reads like a wish for something cleaner, that's the gap Midground is built for: scroll motion and pinned scenes without the widget sprawl (https://midgroundwp.com/ai-agents/agent-ready/).

How do Elementor add-ons and third-party widgets hurt page speed?

Finding the offending widget with the coverage report

Open Chrome DevTools, run the page, and check the Coverage tab under the three-dot menu. Every CSS and JS file an add-on loads shows up with bytes used versus bytes shipped. A widget used once in the footer that drags in 80 KB of unused script is the pattern you're hunting.

Then test per widget: duplicate the page, delete one add-on widget at a time, and re-run PageSpeed Insights. Ten minutes of deletion beats guessing. Elementor also ships asset controls on its side; its performance documentation covers Optimized CSS Loading and Element Caching, which reduce repeat output from widgets.

The add-on audit: keep, replace, remove

Sort every installed add-on into three buckets. Keep what you actually use on live client pages. Replace single-widget add-ons with GenerateBlocks or a native block where the design allows. Remove the rest, then deactivate and delete, not just deactivate.

If the audit keeps failing because the widgets themselves are the point of the design, that's a builder problem, not a settings problem. It's the reason we built Midground: scroll-driven pages without stacking third-party widgets to get them.

What realistic PageSpeed score can an Elementor site achieve?

Set expectations at 70–90 on mobile and 90+ on desktop for a well-tuned Elementor page with a lightweight theme like GeneratePress or Astra underneath. Those ranges are achievable with the fixes in the sections above: trimming widgets, deferring scripts, compressing media, and cutting add-ons you don't need.

Where's the ceiling? A perfect 100 on mobile is rare on Elementor because the builder ships its own CSS and JavaScript on every page, and you can only strip so much before the layout depends on it. If a client demands 95+ mobile, quote them the honest trade-off: either fewer animated sections, or a leaner builder like Bricks Builder, Greenshift, or GenerateBlocks.

Promise the diagnostic process and the fixes, not a number. Scores also shift with hosting and image weight, which are yours to control.

When should you switch away from Elementor for performance reasons?

There is a point of diminishing returns. If you have cleaned up widgets, fixed LCP and CLS, and pruned add-ons, and PageSpeed Insights still flags unused CSS and JS on every new build, further tuning is a tax you pay per project. A HTTP Archive analysis of page weight shows builder-driven pages carrying more CSS and JavaScript than the median page, which matches what you are seeing in your own reports.

Comparing alternatives by the same PageSpeed criteria

Run each candidate through the same test you ran on Elementor: build one landing page, check the DOM size, unused CSS and JS, and LCP in PageSpeed Insights. Greenshift, Bricks Builder, GenerateBlocks and Midground all let you do this before committing a client site. Midground is built for scroll-motion pages specifically, so judge it on whether pinned scenes and parallax ship without the page-weight penalty you would get hand-rolling GSAP.

What switching actually costs a freelancer

You will rebuild patterns, relearn shortcuts, and move slower on the first two projects. That is the honest price. Against it: you stop paying the per-site tuning tax, and you stop dreading the next update.

A worked diagnosis: from slow Elementor page to fixed, step by step

Step 1-3: test, map, fix

Run PageSpeed Insights on the client's slowest page and screenshot every finding. Suppose the report shows a weak LCP, unused JavaScript, and a large DOM. Map each finding to its source: hero slider widget driving LCP, five inactive add-on widgets loading scripts anyway, nested sections inflating the DOM.

Fix in order: replace the slider with a single optimized image, deactivate unused add-ons, flatten nested sections, enable Elementor's improved asset loading and reduced markup settings. One change per re-test so you know what moved the number.

Step 4: re-test and set client expectations

Re-run the test on mobile and desktop, and keep the before/after screenshots for the client. Report the improvements qualitatively: faster load, fewer flagged issues, a calmer report. If scores stay poor after clean settings and minimal widgets, the honest answer is that the page's structure fights the builder, and switching tools beats more tuning.

FAQ

Why is Elementor slow even after optimization?

Because the builder ships its own CSS and JavaScript on every page regardless of your settings, and nested section markup inflates the DOM. Optimization trims the excess but cannot remove the framework itself. If PageSpeed Insights still flags unused CSS and JS after you have pruned widgets and add-ons, the remaining weight is structural.

What is a good PageSpeed score for a WordPress site built with Elementor?

Aim for 70-90 on mobile and 90+ on desktop with a lightweight theme like GeneratePress or Astra underneath. A perfect 100 on mobile is rare because the builder loads its own assets everywhere. Promise clients the diagnostic process and concrete fixes, not a number, since hosting and image weight also move the score.

Is it worth switching from Elementor to a lighter builder for one slow site?

Usually not for one site. Switching means rebuilding patterns and moving slower on your first couple of projects. Switch when every new build still flags unused CSS and JavaScript after proper tuning, because then you are paying a per-project tax. For a single slow page, run the fixes in this guide first.

Does Elementor's built-in optimization fix unused CSS warnings?

Partially. Improved CSS Loading splits the stylesheet so pages load only the CSS their widgets use, and Improved DOM Output flattens wrapper divs. These reduce the warnings but rarely eliminate them, since the builder still ships its core assets on every page. Re-run PageSpeed Insights after enabling both to see what remains.

Do Elementor Pro widgets load slower than free ones?

Not inherently slower, but Pro adds more widget options, and every widget you place carries its own markup and often its own stylesheet. The weight depends on how many widgets a page uses, not which tier they came from. Test per widget with the DevTools Coverage tab and PageSpeed Insights rather than assuming Pro is the problem.