Three in ten UK-visited websites fail Google's phone speed test

Seven in ten pass, and the rest lose shoppers at the door

The comfortable assumption is that phone speed is a solved problem. Phones are fast, broadband is fast, and most sites now come out of a template. The data offers some comfort: a clear majority of the sites people in the UK use on their phones get a passing grade. It also shows a large minority that does not. On a high street, that would be every third shop with a door that sticks.

Loading is the test sites fail most often

Ask people what annoys them about mobile sites and they will often describe the jump: a button that moves just as they tap it. That is the test sites handle best. On phones, 90.2% of UK-visited websites had a good score for layout stability. Only 80.8% had a good score for loading, the lowest of the three tests. In plain terms, about one site in five is still showing a blank or half-drawn page when it should be showing the product.

Responsiveness sits between the two. Google calls the three tests LCP, INP and CLS. For anyone who runs a website, the order matters more than the jargon: if you fix one thing first, fix loading.

The world's busiest websites pass less often

Surely the giants have this sorted. They have engineering teams, budgets and every reason to care.

They pass less often. Of the 9,688 UK-visited websites ranked in the global top 10,000, 66.8% passed on phones. That compares with 70.1% of the 481,431 less popular UK-visited sites. "Popular" here means global popularity among Chrome users, so this ranking is not a UK league table. The data shows the gap but does not explain it. Even so, it should end the idea that copying the biggest names is a safe route to a fast site.

Wix shops pass most often of the six main shop platforms, WooCommerce least

The platform behind an online shop is the last place to look. The assumption is that a site is a site, and speed depends on whoever built it. Among the six online-shop platforms with at least 1,000 UK-visited websites, the spread is wide. Wix eCommerce sites passed on phones most often, at 91.8%, and WooCommerce sites least often, at 60.2%. Shopify and Squarespace Commerce sit close behind Wix. Magento and WooCommerce share the foot of the six, with almost identical pass rates.

Two cautions apply. First, a platform's score also reflects the sites built on it: their size, their themes and the apps added on top. A small shop running a hosted template is a different thing to measure from a large catalogue with years of add-ons. The table shows where pass rates differ. It does not show that moving platform would change any one site's result. Second, the platform is identified from each site's home page, and sites missing from that check are left out.

A WooCommerce shop is not doomed by its platform either: most of them still pass. The difference is that their owners have more to tune themselves.

What we found on our own homepage

In October we put sayu.co.uk through the same kind of check. Google's PageSpeed Insights gave our homepage a mobile performance score of 60, and on its simulated mid-range phone the main banner took 8.6 seconds to appear. Google counts 2.5 seconds or less as good. On desktop the same page scored 96, so the problem only showed on phones.

Our server had the page ready in around 50 milliseconds. The delay came from everything the browser fetched afterwards: about 350 KB of Google Analytics and Google Ads code, fonts and an icon set loaded from other servers, and images up to twice the size they were shown at. Loading the tracking code on the visitor's first tap or scroll, serving the fonts and a trimmed 28 KB icon set from our own domain, and preloading a banner sized for the screen brought the banner time down to between 1.7 and 2.0 seconds. The mobile score went to 99 to 100. It took one afternoon.

One caution if you try the same. PageSpeed Insights is a lab test on a simulated phone, and it is the quickest way to find what is slowing a page. The pass rates in this study come from real visits, and so does the Core Web Vitals report in Google Search Console, which is where you see whether a fix has worked for your customers.

We wrote up all seven changes, and the ones we decided against, in how we got sayu.co.uk to a perfect PageSpeed score.

What this means for UK businesses

Passing is common, so failing puts you behind the majority of sites your customers already use. The busiest sites in the world are no model to copy here. Loading is the test that trips up the most sites, so start there before you touch anything else. If you run your shop on a platform where fewer sites pass, the platform is not a verdict. It just means you have more settings to check yourself.

The shopper on the bus will never know which of the three tests your site failed. They will only remember where their thumb landed.

How we built it

Pass rates come from the Chrome UX Report, Google's public dataset of real-user measurements. It records page loads from Chrome users who share usage statistics, here limited to visits made in the UK in September 2026. It covers websites UK people use, whether or not they are UK-owned, and includes only sites with enough such traffic. Phone data comes mainly from Chrome on Android. iPhone visits through Safari are not included. These are real-user measurements for one month, a snapshot.

A site "passes Core Web Vitals" when, at the 75th percentile of visits, Largest Contentful Paint is 2.5 seconds or less, Cumulative Layout Shift is 0.1 or less and Interaction to Next Paint is 200 milliseconds or less. Sites without enough responsiveness data are judged on the other two tests. That is why the responsiveness figure covers 380,716 sites against 491,119 for the other two.

Popularity tiers are the Chrome UX Report's own global rank buckets. The global top 10,000 and the less popular sites are separate groups, each measured on its own. Platforms were identified from technology detected on each site's home page in the HTTP Archive mobile crawl of the same month, matched by web address. Sites not in that crawl are left out. We report only shop platforms with at least 1,000 UK-visited websites. Pass rates by platform are descriptive: they reflect the mix of sites on each platform, and we made no attempt to control for site size, theme or add-ons. The year-earlier figures in the table below compare different sets of sites, because the list of sites with enough traffic changes from month to month.