Desktop websites passing
56%
Eight points ahead of mobile in the same real-user dataset.
Mobile improvement since 2023
+12 points
Mobile improved four points more than desktop over two years.
Inner mobile pages
+11 points
Their pass rate was higher than the mobile homepage rate.
Research section
Mobile websites are catching up
The mobile pass rate rose faster than the desktop rate from 2023 to 2025.
Three years of real-user data
Mobile improved faster than desktop
Mobile gained 12 percentage points from 2023 to 2025. Desktop gained 8.
The desktop lead narrowed from 12 points to 8 points.
Mobile improved from 36% passing in 2023 to 48% in 2025. Desktop moved from 48% to 56%. That narrowed the desktop lead from 12 percentage points to 8.
This is real-visitor data, so it reflects the devices, connections, and pages people actually used. It does not come from one controlled speed test.
Research section
Loading is the main mobile weakness
Mobile sites were more likely to respond quickly and stay visually stable than to show their main content quickly.
What passes—and what does not
Loading is the weakest mobile vital
A site passes overall only when loading, response, and stability are all good for real visitors.
Loading
Main content appears
Response
Taps and clicks respond
Stability
The page stays in place
All three
The website passes overall
On mobile, 62% of sites showed their main content quickly enough. The response rate was 77%, and 81% kept the page from shifting unexpectedly.
This matters because a page can feel responsive after it appears but still make the visitor wait too long for the first useful content. Looking only at one speed score can miss that difference.
Research section
The homepage was usually the slower page
Inner pages carried less data and passed Core Web Vitals more often on both mobile and desktop.
The homepage penalty
Inner pages were lighter and passed more often
Homepages usually carry more images, promotions, scripts, and tracking than the pages people visit next.
Desktop
+14 points for inner pagesHomepage
47%
passed all three vitals
2.86 MB
median page weight
Inner page
61%
passed all three vitals
2.0 MB
median page weight
Mobile
+11 points for inner pagesHomepage
45%
passed all three vitals
2.56 MB
median page weight
Inner page
56%
passed all three vitals
1.80 MB
median page weight
On mobile, 45% of homepages passed, compared with 56% of inner pages. The desktop gap was even wider: 47% versus 61%.
Homepages often collect everything at once: a large opening image, featured content, promotions, tracking, chat, and several calls to action. This comparison shows why testing only an article or product page can give a misleading picture of the site.
Research section
Lighter homepages passed almost twice as often
The pass-rate difference between small and very large homepages appeared on both desktop and mobile.
Page weight and the pass rate
Under-1 MB homepages passed almost twice as often
This is a strong relationship, not proof that file size alone caused every pass or failure.
Desktop
1.84×Mobile
1.9×Among homepages under 1 MB, 70% of desktop pages and 57% of mobile pages passed. For pages of 5 MB or more, the rates fell to 38% and 30%.
Page weight is not the only cause of speed problems, but unnecessary images, scripts, and fonts create work on every visit. Removing them helps especially on slower phones and connections.
Research section
Core Web Vitals measure three different problems
The names sound technical, but each metric describes a frustration ordinary visitors recognize.
A plain-language guide
The three vitals catch three different frustrations
Google checks the 75th percentile, so at least three quarters of measured visits must meet each target.
Loading
2.5 seconds or less
The main image or heading appears promptly.
Slow pages leave a blank or unfinished-looking first screen.
Response
200 milliseconds or less
A tap, click, or key press gets a quick visible response.
Poor response makes buttons feel broken or delayed.
Stability
0.10 or less
Text, images, and buttons do not jump unexpectedly.
A shifting layout can make someone click the wrong thing.
Loading measures how quickly the main content appears. Response measures how quickly the page reacts after a tap or click. Stability measures whether the layout jumps while someone is trying to read or act.
A site passes only when all three are good for at least 75% of measured visits. One strong metric cannot cancel out a poor one.
Research section
A speed test and real visitors tell different stories
Use real-user data to see whether people have a problem. Use a controlled test to investigate why.
The field section of PageSpeed Insights summarizes eligible Chrome visits from the previous 28 days. Lighthouse runs a fresh controlled test. A green Lighthouse result does not prove that real mobile visitors are also passing.
Check the homepage and the pages that matter most to the business: product pages, articles, checkout, booking, or account screens. The page-type comparison above shows why one tested URL is not enough.
Sources and method
How the numbers are maintained
July 2025 dataset, published 2026
Mobile and desktop trends, individual-vital pass rates, and page-type comparisons.
July 2025 dataset, published 2026
Median page weight and pass rates for lighter and heavier homepages.
These figures are reviewed against each annual Web Almanac release. If a new edition changes a definition, the old value stays in place until the comparison is reviewed rather than being silently mixed with a new measurement.