| Group | Percentage |
|---|---|
| scored under 90 on Google's PageSpeed test, mobile | 67% |
| scored under 90 on both of two runs | 59% |
| in Google's red band, under 50 | 11% |
| had a lab Largest Contentful Paint over 2.5 seconds | 75% |
Source: Digital Royalty analysis, data collected 5 October 2026
Download data
What we found
On 5 October 2026 we ran 132 UK web pages that sell page-speed or Core Web Vitals work through Google PageSpeed Insights, mobile, twice each, from Google’s servers. We tested the page doing the selling: the service page that ranked in Google for searches such as “page speed consultant” and “wordpress speed optimisation service”.
89 of the 132 (67%) scored under 90, the point where Google’s test turns a score green. The median score was 78. Scores move from run to run, so the conservative reading is the 77 of 130 pages with two scores (59%) that came in under 90 on both runs. 14 (11%) were in Google’s red band, under 50.
Largest Contentful Paint, the time the biggest thing on the screen takes to appear, was the most common problem. 99 pages (75%) took more than 2.5 seconds in the lab test, with a median of 4.1 seconds. In all, 105 pages (80%) missed at least one of the three lab Core Web Vitals lines.
Ranking well made little difference. Of the 56 pages in the top 10 for their best search, 35 (62%) scored under 90.
A handful of pages promise a score. At least five offer a guaranteed score or a “90+” result on the page we tested, and all five came in under 90 on both runs. One promises a guaranteed score on mobile and desktop and scored 60 and 50. Another promises 90 or more and scored 37 and 55. We don’t name them, because the gap between the promise and the test is common across the trade rather than the habit of one firm.
Breakdown by type of speed seller
Pages grouped by the search that ranked them best.
| Type of search | Pages | Scored under 90 | Share | Median score |
|---|---|---|---|---|
| Audit and consultant | 15 | 11 | 73% | 84 |
| WordPress and WooCommerce | 35 | 25 | 71% | 67 |
| General speed optimisation, including city searches | 69 | 45 | 65% | 80 |
| Core Web Vitals | 11 | 7 | 64% | 89 |
| Shopify, Magento and Webflow | 2 | 1 | 50% | 85.5 |
WordPress speed sellers had the lowest median score, and audit and consultant pages the highest share under 90. The Shopify, Magento and Webflow row is two pages, so we read nothing into it.
What this does and doesn’t show
Under 90 on this test is not a failed Core Web Vitals assessment. The assessment is based on real-user data, and Google has that for only 3 of the 132 sites, all at whole-site level. Two of those three fail. That is too few to give a rate, so we don’t, and we describe results as “not green on Google’s PageSpeed test”.
A sales page is also not a firm’s client work, and speed agencies could fairly say so. The shop window is still what a buyer reads and what Google ranks, and on 67% of them it wouldn’t pass the test being sold.
We tested ourselves
We sell performance optimisation and technical SEO, so we ran our own pages through the same tool, mobile, two runs each. On 5 October 2026:
| Digital Royalty page | Score (median of two runs) |
|---|---|
| Homepage | 99 (98 and 100) |
| /services/performance-optimisation/ | 99 (99 and 99) |
| /services/technical-seo/ | 99 (99 and 99) |
All three pass every lab line. Before a performance fix on 3 October our homepage scored between 88 and 98 on different runs, which is the same run-to-run variance we describe for everyone else. Google has no real-user data for our site either.
If you want to know where your own site stands, our free website audit runs the same checks and tells you what is slowing the page down.
Method
Method and limits
Data collected 5 October 2026
We found the pages from 96 UK Google searches (68 to the top 30 results and 28 to the top 100), covering speed, Core Web Vitals, WordPress, Shopify, audit and consultant terms, and 25 city variants. Organic results only. That gave 1,516 domains, which we reduced to 857 by filter and then reviewed by hand. A page stayed in if it belonged to a UK business that sells speed, Core Web Vitals or performance work or an audit. We kept one page per domain, the best ranked, which left 133 pages. 132 could be measured.
We ran each page through PageSpeed Insights (v5, mobile) twice, one at a time with a 2 second gap, on 5 October 2026 between 12:28 and 13:40 UTC. The score is the median of the two runs. A lab pass means Largest Contentful Paint of 2.5 seconds or under, Total Blocking Time of 200 ms or under and Cumulative Layout Shift of 0.1 or under.
On the first run of the study (3 October) we re-checked 20 random rows with a third run and a read by hand. 18 of 20 sold speed work (two were removed, along with two blog posts found in a sweep of all 137 candidates), all 20 were UK, and the third run agreed on the score band for 17 of 20.
Limits:
- Lab scores move. The median gap between a page’s two runs was 3 points, but 33 pages moved by 10 or more and 19 straddled 90. Allow about 3 pages either way on the headline share.
- Field data is thin: 3 of 132 sites, no individual pages.
- The study covers the UK, one day, and pages that rank.
- A sales page is not the firm’s client work.
- This is a snapshot. Page speed changes whenever a site does.
- We name no firm on this page. The public download is anonymised.
FAQ
Does under 90 mean a page fails Core Web Vitals?
No. The Core Web Vitals assessment uses real-user data. Our score is a lab test, and 90 is the line where Google’s tool turns the score green.
Why test the sales page and not client work?
It is public, it is the page that ranks, and anyone can re-run the test on it. We have no access to clients’ sites.
Why do scores differ between runs?
Lab tests simulate a mid-range phone on a throttled connection, and results shift with server load and what the page happens to load. That is why we ran each page twice and report a conservative figure alongside the headline.
Can I check a page myself?
Yes. Paste its address into Google’s PageSpeed Insights and select mobile. Expect a few points of variation.
Can I use the data?
Yes. Download the anonymised dataset (CSV): one row per page, with a row id, rank band, type of search, both scores with timestamps and the lab timings. Business names and URLs are removed. Credit Digital Royalty with a link to this page. For the method notes, email press@digitalroyalty.co.uk.