So liest du die Ergebnisse von Google PageSpeed Insights
Der Google Seiten-Speed-Checker zeigt zwei Arten von Ergebnissen. So verstehst du PageSpeed Insights: Felddaten vs. Labordaten, Core Web Vitals, die Punktzahl und was du zuerst beheben solltest.
PageSpeed Insights is Google's free page speed checker, and its report is easy to misread. It shows two different kinds of results on one screen, a score that isn't the same thing as passing Core Web Vitals, and a list of suggestions that aren't equally important. This guide walks through the report from top to bottom, explains what each part means, and shows what to fix first.
The two halves of a PageSpeed Insights report
When you run a URL through PageSpeed Insights, the report has two main sections, and they answer different questions.
- "Discover what your real users are experiencing" is field data. It comes from the Chrome User Experience Report (CrUX): anonymised measurements from real Chrome users who visited the page over the previous 28 days.
- "Diagnose performance issues" is lab data. It comes from Lighthouse loading the page once, right now, on a simulated device and network.
Field data tells you how fast the page actually is for visitors. Lab data tells you why, and is the part you use to debug. When they disagree, field data is the one Google uses for Core Web Vitals.
Reading the field data section
"This URL" vs "Origin"
At the top of the field data you can switch between This URL and Origin. "This URL" shows data for that exact page. "Origin" combines all pages on the site. If a page doesn't have enough traffic of its own, PageSpeed Insights may show only origin data, or no field data at all, with a message that there isn't enough real-user data. That's normal for new or low-traffic pages, and it doesn't mean anything is wrong.
The Core Web Vitals assessment
The headline result is Core Web Vitals Assessment: Passed or Failed. It's based on three metrics, each measured at the 75th percentile of visits, so 3 out of 4 visits must be at least this good:
| Metric | What it measures | Good | Poor |
|---|---|---|---|
| Largest Contentful Paint (LCP) | How quickly the main content appears | 2.5 s or less | Over 4 s |
| Interaction to Next Paint (INP) | How quickly the page responds to taps and clicks | 200 ms or less | Over 500 ms |
| Cumulative Layout Shift (CLS) | How much the layout jumps while loading | 0.1 or less | Over 0.25 |
Anything between "good" and "poor" counts as "needs improvement". The page passes when all three metrics with enough data are in the good range.
Below the three Core Web Vitals you'll also see First Contentful Paint and Time to First Byte. They don't decide the assessment, but they help explain it: a slow Time to First Byte usually means a slow server, and that delays everything else.
Why field data changes slowly
Because field data covers the last 28 days, a fix you made today only shows up gradually as new visits replace old ones. Expect it to take about four weeks to see the full effect. Use the lab section to confirm a fix straight away.
Reading the lab data section
The performance score
The big number in a circle is the Lighthouse performance score, from 0 to 100. It is 90 or more for good, 50 to 89 for needs improvement and under 50 for poor. It's calculated from five lab metrics, weighted like this in current Lighthouse versions:
- Total Blocking Time: 30%
- Largest Contentful Paint: 25%
- Cumulative Layout Shift: 25%
- First Contentful Paint: 10%
- Speed Index: 10%
Two things follow from this. First, Total Blocking Time and LCP together make up more than half the score, so heavy JavaScript and a slow main image hurt the most. Second, the score is a lab estimate, so a page can score 70 and still pass Core Web Vitals in the field, or score 95 and fail them.
Mobile and desktop tabs
PageSpeed Insights shows Mobile by default. The mobile test simulates a mid-range phone on a throttled mobile connection, so the score is almost always lower than desktop. That's expected. Because many visitors browse on phones, the mobile result is usually the one to improve first.
Why the score changes between runs
Run the same page twice and the score can move by a few points even though nothing changed. Server response time, third-party scripts, ads and network conditions all vary. Run the test three times and look at the typical result instead of reacting to a single run.
The suggestions: what to fix first
Under the metrics, PageSpeed Insights lists specific issues, each with an estimate of the time it could save. Newer versions group many of these as insights, alongside diagnostics and a list of passed audits. Some suggestions you'll often see:
- Largest Contentful Paint element: shows which element is the LCP. If it's an image, that image is where to start.
- Properly size images / serve images in modern formats: your images are bigger than they need to be.
- Eliminate render-blocking resources: CSS or JavaScript files that must load before anything shows.
- Reduce unused JavaScript: code that downloads but isn't needed on this page.
- Avoid large layout shifts: elements that move as the page loads, often images without set dimensions or late-loading ads.
- Reduce initial server response time: the server is slow to start sending the page.
Fix in this order: start with whatever affects the metric you're failing in the field. If LCP fails, look at the LCP element and the server response. If INP fails, look at long JavaScript tasks. If CLS fails, look at layout shifts. Then work down the suggestions with the biggest estimated savings. Our guide to website speed optimization walks through the common fixes step by step.
What does "not enough real-world speed data" mean?
It means the Chrome User Experience Report doesn't have enough visits recorded for that page, or for the whole site, to report reliable numbers. New pages, small sites and pages behind a login often show this. It isn't a penalty or an error. Until field data appears, use the lab results to judge speed, and check the Origin view, which pools data across your site and needs less traffic per page.
PageSpeed Insights vs Search Console's Core Web Vitals report
Both use the same real-user data, but they show it differently. PageSpeed Insights reports one URL (or the whole origin) at a time, with lab diagnostics attached. Search Console's Core Web Vitals report groups all of your pages into "good", "needs improvement" and "poor" for mobile and desktop, and lists example URLs for each issue. Use Search Console to find which groups of pages have a problem, then PageSpeed Insights on an example URL to find out why. When you've fixed an issue, Search Console's "Validate fix" button starts a 28-day check of the affected pages.
Common mistakes when reading the report
- Chasing 100. A score in the 90s is "good". The last few points rarely change anything for visitors.
- Ignoring the field data. If the Core Web Vitals assessment passes, your real visitors are having a good experience even if the lab score is lower than you'd like.
- Testing only the homepage. Your product, article or landing page templates may be much slower. Test one of each.
- Comparing different tools. Each tool uses its own device, network and location settings, so compare results from the same tool over time.
Is PageSpeed Insights the same as Lighthouse?
The lab half is Lighthouse. PageSpeed Insights runs Lighthouse on Google's servers and adds the real-user field data from CrUX. Running Lighthouse in Chrome DevTools uses your own computer and connection, so results differ more from run to run. Our Lighthouse vs PageSpeed Insights guide covers the differences.
Does my PageSpeed score affect my Google rankings?
Not directly. Google's ranking systems use Core Web Vitals from field data as part of page experience, not the Lighthouse score. Google also says it still ranks the most relevant content first, so good Core Web Vitals help but rarely outweigh relevance. A faster page still matters for visitors, who leave slow pages before they read anything.
How often should I check PageSpeed Insights?
Check after any change that affects loading, such as a new theme, plugin, script or image-heavy page, and otherwise once a month for your most important pages. For your whole site, the Core Web Vitals report in Google Search Console groups pages by status, so you can spot problems without testing every URL.
Quick ways to check page speed
- One page, with Google's data: run PageSpeed Insights.
- A quick mobile check plus a badge for your site: our free website speed test runs the same PageSpeed Insights test on mobile without signup.
- Comparing tools: see the best free website speed tests.
Related free tools and guides
- Free website speed test — mobile speed and score for any URL
- Lighthouse checker — all four Lighthouse scores
- Website speed optimization: test, fix and re-test