Performance · 15 June 2026
Reading PageSpeed Insights Correctly and Deriving Measures

You enter your address into PageSpeed Insights, click "Analyse," and get a number between 0 and 100, a few red bars and a long list of technical terms. Many close the window again at this point. Understandable, because at first glance the report looks like a diagnostic device whose manual someone has misplaced.
Yet there's an enormous amount that's usable in this free Google tool, once you know what to look at. I'll show you how to read the Lighthouse report, which numbers really count and which measures can be concretely derived from them.
The big number is only the beginning
Right at the top sits the performance score: a tally from 0 to 100. Green from 90, orange in between, red under 50. This number is tempting because it seems so simple, but it's only a summary. It tells you that something is slow, not why.
More importantly: the value fluctuates from measurement to measurement. Run the same test twice in a row and there can be five to ten points of difference between them. So don't chase the perfect 100. A stable value above 85 is perfectly solid for an SME website.
Lab or real users? The decisive difference
PageSpeed Insights shows two kinds of data, and most people confuse them:
- Field data (at the top, titled "Discover what your real users are experiencing"): this is how real visitors experienced your page over the last 28 days. That's the truth that counts.
- Lab data (below it, from Lighthouse): a simulated test run on a standardised device. Useful for diagnosis, but not what your customers actually feel.
If field data is missing, your page simply has too few visitors for reliable statistics, no drama. Then you rely on the lab data and keep in mind: that's a guide value, not a verdict.
The three values that really count
Under the field data you'll find the Core Web Vitals, Google's measure of the user experience. There are three you should know:
- LCP (Largest Contentful Paint): how quickly does the largest visible content appear? Good is under 2.5 seconds.
- INP (Interaction to Next Paint): how smoothly does the page respond to clicks and taps? Good is under 200 milliseconds.
- CLS (Cumulative Layout Shift): does the layout shift while loading? Good is a value under 0.1.
If there's a green bar everywhere here, your page is fine from Google's point of view, no matter what the big score says. If one jumps to orange or red, you know exactly where to start.
From diagnostic report to measure
Scroll further down and you reach the "Diagnostics" section with concrete recommendations. Here it's worth looking closely, because Google sorts by savings potential: at the very top is what brings the most. The most common hints and what they mean:
- "Serve images in modern formats" / "Properly size images": your most common brake. Images uploaded too large, in the wrong format. Solution: dimension correctly and serve as WebP.
- "Reduce unused JavaScript": the page loads code it doesn't need at all, typical of page builders and heavy themes with lots of plugins.
- "Render-blocking resources": certain files block the building of the page. This can be defused technically.
- "Reduce initial server response time": your hosting responds too slowly. Sometimes just a better provider helps.
Each of these points can be expanded and shows exactly which files are affected. You don't have to repair it yourself, but now you know what to talk about with your developer.
What you should concentrate on
The list can be long, and not every point is equally important. My advice: start at the top, with the entries with the biggest savings potential, and work your way down. Three or four targeted improvements often bring more than twenty small ones.
And don't forget to test on mobile. At the top of the report you can switch between "Mobile" and "Desktop." Most visitors come from a phone, and that's exactly where the values are almost always worse. The mobile view is the more honest one.
A number is not an end in itself
PageSpeed Insights is an excellent compass, but not a goal in itself. A green 95 is of little use if visitors still find the page sluggish, and a 78 is no reason to panic if the Core Web Vitals are fine. Read the report as what it is: a concrete to-do list, sorted by impact.
If you're battling through the terms and aren't sure what really counts, I'm happy to go over it with you. Often it's two or three levers that make the big difference. Just get in touch via the contact form, or browse through more posts around fast websites.


