Website speed test

How to use the website speed test

Type a domain or paste the link of the page you want to check, choose Phone or Computer, and press Test speed. Google's test browser opens the page from scratch, measures every moment of the load and sends the result back, usually within 15 to 40 seconds. While it works, the website speed test also asks Chrome's real-user report how fast the same site was for actual visitors over the last 28 days.

The phone test slows the network and the processor down to a mid-range phone, the same way Google measures pages for search. Most visitors are on phones, so start there and test the computer view after.

What the page speed test shows

• A score from 0 to 100. 90 and above is fast, 50 to 89 needs work, under 50 is slow. The score is built from five lab measurements, and the time the page stays frozen by scripts and the main content time count the most.
• The five lab measurements in plain words: when the first thing appears, when the main content shows, how long scripts keep the page frozen, how much the layout jumps, and how quickly the visible page fills in. Each one says what it means for a visitor and whether it is good by Google's limits.
• Real visitors. The Core Web Vitals test uses Chrome's data: loading of the main content, response to taps and clicks, and layout stability. A page passes when all three are good for at least three visits in four. Bars show the share of good, middling and poor visits, and a chart shows the main content time week by week for the last 25 weeks.
• What would help most. Every suggestion from Lighthouse is rewritten as a plain sentence with what to do, sorted by the time it could save. When images are the biggest problem, resizing them and saving them as WebP or AVIF is usually the quickest win.
• Every file the page loads on a time line, with the page weight split into images, scripts, styles and fonts, and a picture of the page as the test browser saw it.

Lab test or real visitors: which one counts?

Both, for different reasons. The lab test runs the same way every time, so it is the one to use while you fix things. The real-user data is what Google uses when it looks at page experience, and it reflects your visitors' phones and networks. A page can do well in the lab and still fail for real visitors on slow connections, or the other way round. When the two disagree, trust the real-user numbers and use the lab test to find out why.

To see how fast your server answers from different countries, open the Uptime section of the same check. It times the first answer from up to 20 places around the world, and the website checker report puts both side by side.

Privacy and limits

The test runs on Google PageSpeed Insights through our server, because Google's service needs a key that must not sit in a web page. Google loads the page you name, so only test public pages. We keep a copy of the result for 15 minutes so that repeated tests of the same page come back at once, and we store nothing after that; the privacy page has the details. A short automatic check keeps bots away, and each visitor can run a limited number of tests an hour, because every test spends a shared Google allowance.

Questions

Why does my score change between tests?

Networks, servers and third-party scripts never behave exactly the same twice. Differences of a few points are normal; run the test two or three times and look at the trend.

Why does Chrome have no real-user data for my site?

Chrome only reports on sites with enough visitors who share usage statistics. New and small sites often have none yet. The lab test still works for them.

Does a faster site rank higher?

Speed is one of many signals Google uses. It rarely beats better content, but a slow page loses visitors before they read anything, so it is worth fixing either way.