What does Google PageSpeed measure?
PageSpeed Insights gives two scores per page: one for mobile and one for desktop. The score runs from 0 to 100, with green (90+), orange (50-89) and red (below 50).
The score is a weighted average of six measurements, where Core Web Vitals weighs most for SEO: Largest Contentful Paint (LCP), Interaction to Next Paint (INP) and Cumulative Layout Shift (CLS).
The three Core Web Vitals
LCP — Largest Contentful Paint: how fast does the largest element on your page appear? Target: under 2.5 seconds. Above 4 seconds: red.
INP — Interaction to Next Paint: how fast does your page respond to a click or keypress? Target: under 200 milliseconds. Above 500 ms: red. (Replaced FID in March 2024.)
CLS — Cumulative Layout Shift: how much does your layout shift during loading? Target: under 0.1. Above 0.25: red.
Only when all three are green does Google consider a page fast.
What does a score of X mean?
90-100: fast page, little room for improvement. Going further isn't worthwhile.
70-89: reasonably fast but there's gain to be made. Invest if competition is faster.
50-69: slow page. Direct impact on bounce and conversion. Optimisation pays off.
Below 50: critical. Visitors leave before the page loads. SEO and conversion both suffer.
A score of 100 isn't a holy grail. 85-95 is practically optimal. Optimising above that yields less than other work.
Real-world example
A Brabant webshop has a PageSpeed score of 32 (mobile). LCP sits at 5.8 seconds due to an unoptimised header image of 4 MB.
After optimisation: header compressed to 180 KB (WebP format), font-loading deferred, critical CSS inline. PageSpeed score: 79. LCP: 1.9 seconds.
Result after 30 days: bounce rate from 68% to 41%. Conversion from 1.4% to 2.7%. Same ad budget, nearly doubled revenue.
Tools to measure PageSpeed
PageSpeed Insights (pagespeed.web.dev): Google's official tool. Uses field data from real users (CrUX) if available, lab data from Lighthouse otherwise.
Search Console > Core Web Vitals: aggregate field-data reporting across your whole site. Best source to see which page groups Google sees as slow.
WebPageTest: granular info per resource. For those who want to know which script is the bottleneck.
Lighthouse in Chrome DevTools: local testing during development. Not representative of real users but useful during fixes.
Common mistakes
Optimising only for desktop. Google ranks mobile-first; mobile score weighs heavier.
Seeing PageSpeed as goal instead of means. Visitors don't book at PageSpeed 95, they book at a fast page they find valuable.
Confusing field-data with lab-data. Lab-data changes after every edit; field-data has a 28-day rolling window. The first reacts immediately, the second slowly.
Optimising without measuring. First identify the heaviest resources (often images and third-party scripts), then fix targeted.
Impact on SEO and ranking
Since 2021, Core Web Vitals are an official ranking factor. Not the largest, but a tiebreaker. With two pages of comparable content, the faster one ranks higher.
Indirect effect is larger: faster pages have lower bounce rate, longer session duration and more pages per session. Google notices this via behaviour patterns.
For SMEs with local ranking ambition: ensure you score in the green on all three Core Web Vitals. Optimising above that hardly brings more ranking.
Want to know if your site loads fast enough to rank and convert?
Request a performance scan →