Why website speed is measured on mobile

PageSpeed Insights loads the page in a simulated setting: a mid-range phone, a mobile network and one location. According to Google's documentation, the mobile test simulates a Moto G4-type device on a mobile connection. The desktop test uses a simulated computer with a wired connection.

The choice is deliberate. A top-end phone on fast Wi-Fi would make almost every page look fast. A modest device on a slower connection shows where the page needs work: large images, unneeded code, a layout that jumps while loading. The PageSpeed Insights documentation says the tool gives results for mobile and for desktop. The mobile score is therefore a choice of viewpoint, the more demanding of the two.

Google looks at the mobile version first because it indexes and ranks sites mainly on the mobile version of their content. The Google Search documentation says the mobile version of a site is used, as read by the smartphone agent. So the mobile page is the one that counts.

This matters in practice for speed. When the mobile version is just as fast and complete as the desktop version, Google sees your site at its best. Google recommends responsive design, because one page serves both screens. The documentation says which version is examined. It does not promise places in the results.

Lab data comes from a controlled test: one device, one network, one location, whenever you ask for it. Field data shows the experience of real visitors over time. The first helps to find and fix problems. The second shows how visitors were actually served. It is normal for them to give different results.

The web.dev article on the differences between lab and field data explains it simply: a lab test is one device, on one network, from one location, while field data comes from everyone who visits the page. It suggests using field data to set priorities and lab data to track down errors. PageSpeed Insights shows both on the same screen. That is why a page can do well on one and need work on the other.

The score changes because speed also depends on things outside the page: the server's response at that moment, the route through the network, ads or tests running on the page, the machine doing the measuring. Lighthouse, the engine behind the lab score, says much of this variation is not caused by Lighthouse itself.

The performance scoring guide lists typical causes: changes in the ads that appear, changes in how traffic is routed across the internet, tests on different devices. The right approach is to measure several times and keep the median. This is also how the speed guarantee we give works: five measurements, and the median counts.

Core Web Vitals are three measurements of real experience: how fast the main content appears, how fast the page responds to a tap, and how stable the layout stays. Google suggests judging them at the 75th percentile of page loads, separately for mobile and desktop, so that the slower visits show too.

The web.dev guide to Web Vitals gives clear targets: main content within 2.5 seconds, a response to an action within 200 milliseconds, and a layout shift score of 0.1 or less. Because mobile and desktop are measured separately, a site can be fast on desktop and slower on mobile. Putting mobile first makes sure you see the full picture.

Check your important pages on mobile, a few times, and read the result as a direction for improvement. One measurement on the fastest laptop you have does not show your visitors' experience. The steps below take about ten minutes and need no technical knowledge.

1. Open PageSpeed Insights and enter the address of your home page.

2. Choose the mobile tab, not desktop.

3. Run the test three to five times and note the Performance score each time.

4. Keep the median value, not the best one.

5. Repeat for one page of each main type: a service page, an article, the contact page.

6. Read the list of suggested improvements and give it to whoever maintains the site.

Google's documentation calls 90 and above good, 50 to 89 a score that needs improvement, and below 50 poor. If you prefer us to take it on, the free website check looks at what affects how understandable and trustworthy a website is.

A speed guarantee covers a measurable result in a specific test, not a place in search results. Ours is a Performance score of at least 80 in PageSpeed Insights, on mobile, on the day of delivery, for the home page and one page of each main type, as the median of five measurements. If we do not reach it, we fix it at no charge.

We wrote the rules this precisely so you know exactly what you get. You will find all the details on the speed guarantee page. The guarantee covers the score, not places on Google or sales, which depend on many other factors. Measuring on mobile makes the score reliable, because it is the most demanding test.