How to Check Your Website Speed and What the Numbers Mean

The connection in your office is the worst place to judge your website's speed, because it is usually far quicker than the phone signal your customers rely on. When a customer says the site crawled on the bus while it opens instantly at your desk, the real question is how to check website speed the way they experienced it. Google's own tools answer that, each test takes about a minute, and none of them costs anything. Running the test is the easy part. Reading it is harder, because the result arrives as a score, a row of coloured numbers with unfamiliar names and a list of technical suggestions. Once you can read those numbers, a report will tell you plainly whether you have a real problem, a small one or none at all.

Before you test: choose the right pages and the right device

Speed is measured one page at a time rather than for a whole website, so start by choosing which pages to test. Your home page is the obvious first choice, followed by the page most visitors land on from Google, which for many businesses is their main service page. Add one page heavy with photographs, such as a gallery, a menu or a portfolio, because large images are where many small sites slow down. Always test the published site that customers see, not the preview inside your website editor. Test the mobile version first, because Google mainly uses the mobile version of your pages when it decides how to rank them. Speed is one part of a wider health check, and our guide on how to check your website SEO walks through the rest of it in about half an hour.

It also helps to know whether speed is your biggest weakness or one of several before you spend an afternoon on it. Running a free SEO audit of your website gives your site a score out of 100, checks every public page and shows the searches you could win, so you can see where speed sits among everything else. Then, when you test speed itself, expect the results to move a little from one run to the next, because the test is affected by network conditions and server load at that moment. Run each page two or three times and take the typical result rather than the best or the worst. Write down the date and the main numbers each time. That simple record will show you later whether your changes have made a difference.

How to check website speed with PageSpeed Insights

PageSpeed Insights is Google's main speed test, and you can find it by searching its name on Google. Paste the full address of the page you want to test into the box, including the https at the start, and press the button to analyse it. After a short wait, the report appears with two tabs, one for mobile and one for desktop. Start with mobile, since that is where most problems show up and where Google looks first. The report is long, but it is divided into two clear parts that answer two different questions. Understanding the difference between those parts is the key to reading everything else correctly.

The top part describes what real visitors have experienced. It uses data collected from people visiting the page in Chrome over the previous 28 days, and it shows whether the page passes or fails what Google calls the Core Web Vitals assessment. You can switch between figures for the single page you tested and figures for your whole site. Many small business sites see a message saying there is not enough data here, simply because they do not yet have enough visitors for Google to report on. That is normal and not a sign of a problem. When it happens, the lower part of the report becomes your main guide.

The lower part is a controlled test, run once on the spot. Google loads the page while imitating a mid-range phone on a slowish mobile connection, then reports a performance score out of 100 and five measurements of how the page loaded. Below those come suggestions for improvement, often with an estimate of the time each could save, and a set of diagnostic notes. The two parts can disagree, and when they do, the real visitor data is the better judge of how your site actually performs. The controlled test is more useful for finding the causes, because it shows exactly what happened during one load. Read the top for the verdict and the bottom for the reasons.

What the numbers mean

The performance score

The performance score runs from 0 to 100 and is coloured to make it quick to read: 90 to 100 is green and counts as good, 50 to 89 is orange and needs improvement, and anything below 50 is red and poor. It is a weighted blend of the measurements from the controlled test, not a separate test of its own. Mobile scores are usually lower than desktop scores for the same page, because the mobile test imitates a slower phone and connection on purpose. Do not chase a perfect 100, which is hard to reach for any page with photographs and is rarely worth the effort. A page in the green or the upper orange whose real visitor data passes is in good shape. A red mobile score, on the other hand, is worth acting on soon.

Largest Contentful Paint

Largest Contentful Paint measures how long it takes for the biggest piece of content in the first screen to appear, which is usually the main photograph or the main heading. Google counts 2.5 seconds or less as good and anything over 4 seconds as poor. This is the number that most closely matches a visitor's sense of whether a page has loaded. The usual cause of a slow reading on a small business site is a large photograph at the top of the page, often uploaded straight from a camera or phone at many times the size the screen needs. A boutique hotel in Zanzibar with a full-width beach photograph at the top of every page is a typical example. Shrinking and compressing that one image can transform the reading.

Interaction to Next Paint

Interaction to Next Paint measures how quickly a page responds when someone taps a button, opens a menu or types into a form. Google counts 200 milliseconds or less as good and anything over 500 milliseconds as poor. Because it depends on real taps and clicks, it only appears in the real visitor data at the top of the report. In the controlled test, a measurement called Total Blocking Time stands in for it, showing how long scripts kept the page too busy to respond while it loaded. Heavy scripts are the usual cause, especially chat widgets, social media feeds, tracking codes and elaborate menus added over the years. If this number is poor, look first at the extras your page loads rather than at its text or images.

Cumulative Layout Shift

Cumulative Layout Shift measures how much the content jumps around while the page loads, and it is a score rather than a time. Google counts 0.1 or less as good and anything above 0.25 as poor. A high reading means things are moving under the visitor's finger, which is how people end up tapping the wrong button. Picture a restaurant in Kuala Lumpur whose booking button slides down the screen as the menu photograph loads above it, so a diner taps the phone number by mistake. Common causes are images without set dimensions, cookie notices or banners that push content down, and fonts that change size as they load. Fixing the cause is usually a small, quick change once you know which element is moving.

The supporting numbers

The report shows a few more measurements that help explain the headline ones. First Contentful Paint is the moment the first text or image appears, and Google counts 1.8 seconds or less as good. Speed Index describes how quickly the visible part of the page fills in overall. Time to First Byte, shown in the real visitor data, measures how long the server takes to start sending the page, and Google counts 0.8 seconds or less as good. A slow first byte points to the server or the way the page is built rather than to your images, which is useful to know before you spend time compressing photographs. Read these numbers as clues rather than goals, because they explain why the main measurements look the way they do.

Other ways to check your speed

Google Search Console offers a wider view through its Core Web Vitals report, which you will find under the experience section once your site is verified. Instead of testing one page, it groups your pages by how they perform for real visitors on mobile and desktop, sorting them into poor, needing improvement and good. Like the top part of PageSpeed Insights, it only shows data when your site has enough visitors, so a very new site may see little at first. Its great strength is showing patterns, such as every gallery page sharing the same problem. Once you have fixed an issue, you can ask Search Console to validate the fix, and it will monitor the affected pages over the following 28 days. That makes it the best place to confirm that a change has really helped.

The most revealing test of all needs no tool. Borrow an older phone, switch off the wireless connection so it uses mobile data, open a private browser window and search for your business on Google, then tap through to your site. Count how long it takes before you could comfortably read the page and tap a button. Try it somewhere your customers actually are, such as a busy market, a train or a suburb with weaker signal. If you are comfortable with more technical tools, the developer tools inside Chrome include the same controlled test as PageSpeed Insights, which is handy for testing a page before it goes public. The phone test, though, is the one that tells you how your customers feel.

Turning the numbers into a short fix list

Once you have read the report, turn it into a short list of fixes, starting with whatever affects the worst number. The suggestions section of PageSpeed Insights helps, because it shows which changes could save the most time. On most small business sites, images come first: resize each one to the largest size it will actually be shown at, compress it and use a modern format such as WebP where your website allows. Make sure images further down the page only load as the visitor scrolls towards them. Next, look at the extras, such as chat widgets, embedded social feeds, video backgrounds, image sliders and additional fonts, and remove any that do not earn their place. Each one you remove makes the page lighter for every visitor.

If the first byte is slow on every page you test, the problem lies with the server rather than the pages, and that is a conversation for whoever looks after your site's hosting. After each change, run the same tests on the same pages and compare them with your earlier record. Remember that the real visitor data covers the previous 28 days, so it will take a few weeks for improvements to show at the top of the report, even when the controlled test improves at once. Speed is one of several things worth checking regularly, and our AI SEO Tools can help with the others, from finding the searches your pages could win to planning your marketing for the months ahead. Return to your speed tests every few months and after any major change to your site. Small slowdowns creep in quietly, and a regular check catches them early.

Checking speed is quick once you know what you are looking at. Test the right pages on mobile first, read the real visitor data for the verdict and the controlled test for the causes, and judge each number against Google's own thresholds rather than chasing a perfect score. Most small business sites need only a handful of fixes, usually starting with oversized images and unnecessary extras. Keep a simple record so you can see your progress, and give the real visitor data a few weeks to catch up. Then try the phone test again on the bus, where your customer first noticed the problem. If the page opens before the next stop, you have done what matters.