Does page speed affect SEO? What Google says and what to fix

Is page speed a Google ranking factor?

Yes, through Core Web Vitals, but it is a light one. Google's page experience guide says “Core Web Vitals are used by our ranking systems”, then adds that “There is no single signal” for page experience and that “Google Search always seeks to show the most relevant content, even if the page experience is sub-par.”

So speed doesn't beat a better answer. It matters where answers are close: “for many queries, there is lots of helpful content available. Having a great page experience can contribute to success in Search, in such cases.” If the pages above you answer the search better, start with your content. If yours is one of several equally good answers, speed becomes worth the work.

Two more lines from the same guide save a lot of wasted effort. Good results in Search Console's Core Web Vitals report or other tools “doesn't guarantee that your pages will rank at the top of Google Search results”, and “trying to get a perfect score just for SEO reasons may not be the best use of your time.”

How has Google's use of speed changed since 2010?

Google has used speed in ranking since 2010, and each change came with the same message: relevance outweighs speed.

Each row comes from Google: the 2010 announcement, the 2018 Speed Update post, the 2021 rollout post and the INP change. The under 1% figure describes 2010, not today. And if you read that a 2026 core update tightened the thresholds, check Google's Core Web Vitals page: it still lists 2.5 seconds, 200 milliseconds and 0.1.

What do LCP, INP and CLS measure?

Core Web Vitals are three measurements of how a page feels to a real visitor: how fast the main content appears, how quickly the page responds to a tap, and how much it jumps while loading. Google's targets come from its Core Web Vitals page.

A page passes when all three meet their targets at the 75th percentile, so at least three in four visits must be good. web.dev recommends measuring “the 75th percentile of page loads, segmented across mobile and desktop devices”, so mobile and desktop are judged separately.

What is CLS in SEO?

CLS is the Core Web Vital for visual stability: it scores how much visible content moves while the page loads. web.dev lists the most common causes as images without dimensions; ads, embeds and iframes without dimensions; content injected after the page has drawn; and web fonts. It counts in the assessment like LCP and INP, so a page that jumps can fail even when it loads fast.

What is the LCP element?

It is the element LCP times: the largest image, text block or video visible on the first screen, often a hero image or the main heading. PageSpeed Insights names it in its report, and it is the first thing to check, because it can change from one test to the next. On our own home page it did, in runs only minutes apart (see below).

Field data or lab data: which one does Google use?

Field data. Chrome's CrUX documentation says the data it collects from real Chrome users “is used by Google Search to inform the page experience ranking factor”. PageSpeed Insights shows that field data first, covering real visitors' experiences “over the previous 28-day collection period”, and Search Console's Core Web Vitals report is built from the same kind of real-world data.

Below the field data, PageSpeed Insights runs a lab test with Lighthouse: “a simulated load of a page on a single device and fixed set of network conditions”. The 0 to 100 score comes from that lab test. Use it to find what to fix, and the field data to judge where you stand.

The lab settings explain why mobile scores run so much lower than desktop. Google says Lighthouse “simulates the page load conditions of a mid-tier device … on a mobile network for mobile, and an emulated-desktop with a wired connection for desktop”. Lighthouse's throttling settings put that mobile network at 150 milliseconds of latency and 1.6 megabits per second, and slow the processor four times. Everything the page needs before it can show its main content takes longer, and the score weights put 30% on Total Blocking Time, 25% each on LCP and CLS, and 10% each on First Contentful Paint and Speed Index. What that means for enquiries on a phone is in our guide to mobile CRO.

Is Google PageSpeed accurate?

It is an accurate record of one simulated visit, not a forecast of your rankings. The lab score moves from run to run, which is why Lighthouse's own documentation says “The median Lighthouse score of 5 runs is twice as stable as 1 run.” The field section is the one that reflects real visitors.

PageSpeed Insights or Lighthouse: what's the difference?

PageSpeed Insights runs Lighthouse in a Google data center and adds field data from the Chrome UX Report when there is enough of it. Lighthouse on its own, in Chrome DevTools or from the command line, runs only the lab test, on your computer and connection, so its numbers won't match PageSpeed Insights exactly, even for the same page.

What if PageSpeed Insights shows no field data?

Then Chrome doesn't have enough real visits to report yet, which is normal for new and small sites. The Chrome UX Report only includes pages and sites that are publicly discoverable and “sufficiently popular”, and Google says “An exact number is not disclosed”. Search Console's Core Web Vitals report shows “No data available” in the same situation. Until data arrives, the lab test is your guide to what to fix, though it isn't the data Google Search uses.

How do you check and benchmark your site's speed?

Start with real visitors, then use the lab to find causes. Five checks, in this order:

1. Open Search Console's Core Web Vitals report. It groups similar pages, shows mobile and desktop separately, and gives each group the status of its worst metric. After a fix, Start Tracking begins “a 28-day monitoring session” to confirm it.
2. Run your key pages through PageSpeed Insights. Read the field data at the top before the score, and note which element it names as the LCP element.
3. Repeat the lab test and keep the median of five runs, with the date, device and connection noted. One run is not a result.
4. Benchmark against the pages that outrank you for your main search, with the same tool and settings. Compare mobile with mobile, and field data with field data.
5. Check whether speed costs you enquiries: compare enquiry rates by device for the pages that fail. To see Core Web Vitals beside enquiries in Google Analytics 4, web.dev's guide sends them as events with Google's web-vitals library.

If your pages pass and enquiries are still low, speed isn't the leak. The checks for that are in our guide to website traffic but no leads.

What makes a website slow, and what should you fix first?

Most slow pages are slow for a handful of reasons, and each hurts a particular metric. Fix the pages that earn search traffic and enquiries first, change one thing at a time, and measure again after each change.

web.dev's guides to LCP and CLS cover each fix in detail. Its warning on images is worth repeating: “Never lazy-load your LCP image”.

How do you eliminate render-blocking resources?

Lighthouse flags two kinds: a script in the page's head without a defer or async attribute, and a stylesheet the browser must download before it draws anything. Its documentation recommends “inlining critical resources, deferring non-critical resources, and removing anything unused”. From Lighthouse 13, the check appears as an insight called Render-blocking requests.

On our own site in October 2026, we removed one render-blocking stylesheet by moving our fonts into the main CSS file, cut about 72 KB of style rules that nothing on the site used (482 KB to 411 KB before compression), loaded the menus on first use and smooth scrolling only for mouse and trackpad users, and re-encoded a background texture from 25.8 KB to 16.4 KB without changing a pixel. Every change was checked against screenshots of every page, at phone and desktop widths.

Does a faster website get more leads?

Sometimes, which is why it is worth measuring rather than assuming. In Vodafone's A/B test, reported by web.dev, “a 31% improvement in LCP led to 8% more sales, a 15% improvement in their lead to visit rate”. In Deloitte's 2020 study of 37 brands' mobile sites, faster lead generation sites saw more visitors get through the form, yet overall conversion rates fell. It is a correlation study using lab data.

So treat speed as one check among several, and compare enquiry rates by device before and after a change. Our guide to why traffic isn't turning into leads sets out that evidence in full.

Page speed myths, and the facts

Questions people ask about page speed and SEO

How important is page speed for SEO?

It matters, but less than relevance. Google says Core Web Vitals are used by its ranking systems, and that it still shows the most relevant content even when the page experience is poor. Speed helps most when several pages answer a search equally well.

Why is the mobile performance score harder to increase in PageSpeed Insights?

Because the mobile test simulates a mid-range phone on a slow mobile connection, with the processor slowed four times, while the desktop test uses a wired connection. Everything the page needs before it can show its main content takes longer, so LCP and the score fall furthest on mobile. Real visitors' data in the field section matters more than the lab score.

How do I speed up my website?

Start with the pages that bring search traffic and enquiries. Right-size and compress the main image and never lazy-load it, defer scripts the first screen doesn't need, remove unused CSS and third-party code, and give images and embeds set sizes. Measure again after each change.

How do I increase mobile page speed?

Cut the JavaScript and CSS the page loads before it can show its main content, because the mobile test slows the processor four times on a slow connection. Serve images at phone sizes in WebP or AVIF, keep the first screen light, and judge the result on mobile field data, not only the lab score.

What affects website speed?

Mainly the size of images, how much JavaScript and CSS the page loads before it can show content, the number of third-party scripts, and how fast the server answers. Web fonts and late-loading banners also make the page jump, which counts against CLS.

What is a good page load time?

Google's Search documentation doesn't give a target for total load time. Its target is for the main content: Largest Contentful Paint within 2.5 seconds for at least three in four visits, with INP under 200 milliseconds and CLS under 0.1.

How do you benchmark a website's speed?

Compare your key pages with the pages that rank above you for the same search, using the same tool and settings. Compare field data with field data and mobile with mobile, and take the median of several lab runs rather than one.

Start with your own data: Search Console's Core Web Vitals report and PageSpeed Insights are free. If speed turns out not to be the problem, see why traffic isn't turning into leads. If you'd rather we find and fix it, speed and Core Web Vitals are part of our technical SEO work, and our fixed-price SEO Audit includes them.

About this article

Google's statements are quoted from its own documentation and announcements, each read on 20 October 2026, with Lighthouse's settings taken from its own documentation and code. The figures for our home page are ten Lighthouse 12.8 runs we made on 20 October 2026: lab data from our own computer, not real visitors' data. The Vodafone and Deloitte findings are their own published results. We have no client speed results to show, so this article claims none. For how page experience fits Google's AI features, see our guide to ranking in AI Overviews.