Short answer
PageSpeed Insights has moved to a new Lighthouse version that updates its loading simulation. A page's lab score can therefore rise or fall even though nothing on the page changed. Problems were also reported in the Google data some rank trackers collect. Recording the tool and version in each report avoids false claims of success.
Highlights
- PageSpeed Insights now runs Lighthouse 13.6.0; JavaScript-heavy sites may feel the difference more.
- Scores measured with different versions are not comparable; rely on real-user (field) data.
- Name the tool and method in rank reports, and cross-check the same pages in Search Console.

2 min readAuthor: UNIT Journal EditorEditor-in-chief: Uğur Deniz İlhan
If the PageSpeed score in your monthly report suddenly moved, the cause may not be a change on your site. According to Search Engine Journal's weekly roundup, PageSpeed Insights now runs Lighthouse 13.6.0, which updates its loading simulation.
Why does the same page get a different score?
Because the lab score is a simulation and the simulation's rules changed. According to Google's Barry Pollard, the new version gives more accurate estimates of First Contentful Paint (FCP) and Largest Contentful Paint (LCP), and JavaScript-heavy sites may feel the difference more. The same report reminds readers that field data, collected from real users, is separate from this simulation.
What should rank reports watch for?
The roundup also names two issues in the Google data that rank trackers collect: unrelated one-word listings can push a tracked page out of the collected results, and unresolved redirect links can hide which page ranked. In SE Ranking's review of 10,000 keywords, 84% of the flagged cases were unrelated one-word results. The theme of the roundup is the same: check the method before the number.
What should sites and agencies do?
- Note the Lighthouse version in every PageSpeed report; do not present lab scores from two versions as an improvement.
- Verify any improvement claim against Core Web Vitals field data and real clicks and impressions in Search Console first.
- Name the tool and measurement method in rank reports; after a sudden drop, cross-check the same query in Search Console.
- On JavaScript-heavy pages, review heavy scripts and render-blocking resources separately.
Explaining a measurement-version change to management correctly is the basis of CRO and analytics work that reads conversion and user behaviour data together. If site speed needs work, you can also re-assess your website and e-commerce infrastructure against the new measurement.
Frequently asked
- If my PageSpeed score dropped, did my site get slower?
- Not necessarily. Because the simulation version changed, the score can move even when the page is the same. First compare the Lighthouse version in your reports and real-user data.
Sources
- Search Engine Journal ·
Follow UNIT Journal
What's new in search, AI and technology, in your feed every day.


