Website Speed Test with Google own engine
Give it a URL, Lighthouse runs against it, and you get four scores, the three core page-experience metrics and a list of improvement opportunities ordered by impact. You can run mobile and desktop together.
- LCP, CLS and TBT
- Mobile and desktop
- Opportunities ranked by impact
Run the tool right here
The official Google score for speed, SEO, accessibility, and best practices
- What it costs $0.012
- Data source DataForSEO
The short answer
A website speed test means running Lighthouse, the same engine PageSpeed Insights sits on, to see how long your page takes to become usable. The three numbers that actually matter are LCP, CLS and TBT, not the big coloured score at the top.
That zero to one hundred score is a weighted average of several metrics, which is exactly what makes it misleading. Two sites can land on the same score while one has an oversized image problem and the other a heavy JavaScript problem, and the fixes have nothing in common. If you intend to act, look at the three metrics under the score rather than the score itself.
LCP says when the largest thing on the first screen appeared, and it is almost always an image or a large block of text. CLS says how much things moved after loading, the experience that makes you click the wrong button; its main source is an image with no width and height. TBT says how long the browser was busy running JavaScript and not answering your click.
The order in which you tackle those three matters, and it is the reverse of what usually happens. Fix CLS first, because it is the cheapest and usually comes down to putting dimensions on images and reserving space for banners. Then LCP, which most often means shrinking and optimising the large image at the top. Leave TBT for last, because it usually means removing or deferring scripts, and that is where something can break.
What you get back
What each output column actually means. Read this before your first run so you do not misread a number.
- LCP (Largest Contentful Paint)
- The render time of the page main content element; green under 2.5s, yellow up to 4s, red beyond that.
- CLS (Cumulative Layout Shift)
- The amount of unexpected element movement during load; green under 0.1, yellow up to 0.25, red beyond that.
- TBT
- The INP field metric requires real Google user data and cannot be measured in a lab test; instead of INP we show Total Blocking Time, the closest lab-based substitute for page responsiveness.
Where the data comes from
Lighthouse is a lab test, not a field test. It runs on a server with simulated network conditions; its score differs somewhat from what real Google users (CrUX data) experience, and it can naturally fluctuate a few points between consecutive runs.
When it helps
Before and after technical site changes (theme change, adding a plugin, image optimization) to measure their effect on speed.
Where it does not help
Lighthouse is a lab test, not a field test. It runs on a server with simulated network conditions, and what it measures is not necessarily what your real visitor experiences on their own phone and connection. Two consecutive runs against the same URL also vary by a few points naturally, so a three or four point difference between runs means nothing and chasing it wastes time. What Google uses for ranking is field data from real users, and you see that in Search Console, not here. This tool is for finding the cause, not for grading.
One thing we learned using it
The temptation almost every site falls into is raising the score instead of making the site fast, and those are not the same thing. Deferring every script until after user interaction lifts the score while that same user now waits longer for buttons to respond. If you have to choose between a higher score and a page that becomes usable sooner, choose the second.
What it costs
No monthly subscription. You top up a wallet and the price of a run is printed on the button before you press it; if you never run it, nothing is deducted.
Questions about this tool
What score is good?
Rather than targeting a score, look at Google own thresholds on the three metrics. Those are the numbers Google has defined good, needs improvement and poor bands for, and they are colour coded in this result. A page with all three metrics in the good band is healthy even with a middling overall score.
Why is my mobile score so much lower than desktop?
Because the mobile test deliberately simulates a weaker device and a slower network. That is normal and every site shows the same gap. If you want a meaningful comparison, compare mobile with mobile, not with desktop.
How much does site speed affect rankings?
It has an effect, but not the one it is advertised to have. Our experience is that speeding up a page on its own usually moves the ranking a little, while it improves bounce rate noticeably. Fix speed for the visitor rather than for Google; the SEO benefit follows on its own.