How to Test Your Website Speed (And Understand the Results)

You want to know if your website is fast or slow. Simple question, right? You Google "website speed test," click on one of the many free tools, punch in your URL, and get a number. Green means good, red means bad. Done.

Except it's not that simple. The number you get depends on which tool you use, whether it's testing mobile or desktop, what network conditions it's simulating, and whether it's using lab data or real user data. Two different tools can give you two very different scores for the same website, and both can be correct.

Understanding how to test your website speed properly, and more importantly, how to read the results, is the difference between knowing what's actually wrong and chasing a number that doesn't matter.

The Tools You Should Use

There are dozens of website speed testing tools out there. Most of them are fine. But there's one that matters more than all the others combined.

Google PageSpeed Insights (pagespeed.web.dev)

This is the one that actually matters. Why? Because Google uses this data to rank your website. PageSpeed Insights runs Google's Lighthouse audit and combines it with real-world data from the Chrome User Experience Report (CrUX). It gives you both simulated lab data and actual field data from real Chrome users.

It's free, it takes 30 seconds, and it's the tool Google themselves recommend. If you only use one tool, use this one.

Lighthouse (in Chrome DevTools)

Lighthouse is the engine behind PageSpeed Insights. You can run it directly in Chrome by opening DevTools (F12), going to the "Lighthouse" tab, and clicking "Analyze page load." This is useful if you want to test a staging site or a page that's not publicly accessible yet.

GTmetrix

GTmetrix is a popular alternative that provides a more visual breakdown of your page's waterfall (how each resource loads over time). It's useful for deeper analysis but uses Lighthouse under the hood, so the core scores will be similar to PageSpeed Insights.

WebPageTest

WebPageTest is the most detailed testing tool available. It lets you choose specific locations, devices, and network conditions. It provides a filmstrip view of how your page renders over time. It's powerful but complex. Use this when you need to dig into specific performance issues, not for a quick check.

Understanding Your Scores

When you run a test in PageSpeed Insights, you get a wall of data. Here's how to read it without getting overwhelmed.

The Performance Score (0-100)

This is the big number at the top. It's a weighted average of several metrics, combined into a single score from 0 to 100.

• 90-100 (Green): Good. Your site is performing well.
• 50-89 (Orange): Needs improvement. There are real issues affecting user experience.
• 0-49 (Red): Poor. Your site is slow enough to actively harm your business.

Don't obsess over getting a perfect 100. Going from 70 to 90 is meaningful. Going from 90 to 100 is diminishing returns. Read our guide on what a good Lighthouse score actually is for more detail on setting realistic targets.

Core Web Vitals (The Three Numbers That Matter Most)

Below the performance score, you'll see Core Web Vitals. These are the three metrics Google uses as ranking signals. They're more important than the overall score.

LCP (Largest Contentful Paint) - How long until the main content of your page appears. This is usually your hero image or main headline.

• Under 2.5 seconds: Good
• 2.5-4 seconds: Needs improvement
• Over 4 seconds: Poor

If your LCP is slow, the user is staring at a blank or partially loaded screen. This is the metric that most directly affects bounce rate. Read our LCP fix guide for specific solutions.

CLS (Cumulative Layout Shift) - How much your page jumps around while loading.

• Under 0.1: Good
• 0.1-0.25: Needs improvement
• Over 0.25: Poor

If your CLS is high, elements are moving after the page appears. The user tries to tap something and it shifts. This is frustrating and makes your site feel broken. Read our CLS fix guide for fixes.

INP (Interaction to Next Paint) - How quickly your page responds when a user taps or clicks something.

• Under 200 milliseconds: Good
• 200-500 milliseconds: Needs improvement
• Over 500 milliseconds: Poor

If your INP is high, the page feels sluggish when interacted with. Buttons take a moment to respond. Menus are slow to open. This is the newest Core Web Vital and the one most sites struggle with. Read our INP fix guide for what to do about it.

Lab Data vs Field Data

PageSpeed Insights shows you two types of data, and understanding the difference is important.

Lab Data is simulated. Lighthouse runs your page in a controlled environment with a simulated mobile device, throttled CPU, and throttled network. It's consistent and reproducible. Every time you run the test, you get roughly the same result (with some variation).

Field Data is real. It comes from the Chrome User Experience Report, which collects actual performance data from real Chrome users who visited your site. This is what Google uses for ranking signals.

Here's the thing: they can tell very different stories. Your Lab Data might show a performance score of 85, but your Field Data might show LCP failures because real users on real mobile networks in real conditions are having a worse experience than the simulation predicted.

When they disagree, Field Data wins. That's what Google uses for rankings. But Field Data has a limitation: it only appears if your site gets enough traffic for Chrome to collect a meaningful sample. If your site is new or low-traffic, you'll only see Lab Data.

The Diagnostics Section

Below the scores, PageSpeed Insights gives you a list of specific opportunities and diagnostics. This is where the actionable stuff lives.

Opportunities are things that, if fixed, would improve your load time. Common ones include:

• "Serve images in next-gen formats" (use WebP instead of JPEG/PNG)
• "Eliminate render-blocking resources" (CSS and JavaScript that blocks the first paint)
• "Properly size images" (serving images that are too large for their display area)
• "Minify CSS" and "Minify JavaScript" (removing unnecessary characters from code)
• "Reduce unused CSS" and "Reduce unused JavaScript" (code that loads but is never used)

Diagnostics are things that don't necessarily affect load time but affect user experience:

• "Avoid large layout shifts" (elements moving after the page loads)
• "Avoid long main-thread tasks" (JavaScript that blocks interaction)
• "Minimize main-thread work" (too much processing happening on the main thread)

Each item tells you the potential savings in milliseconds. Sort by potential savings and start at the top. The items with the biggest potential impact are usually images and render-blocking resources.

Common Mistakes When Testing

Testing only on desktop. Most of your visitors are on mobile. Always test mobile first. In PageSpeed Insights, the mobile tab is selected by default. Keep it that way.

Testing once and panicking. Lab Data can vary between runs. Run the test 3 times and look at the median. A single bad score might be a fluke. Three bad scores is a real problem.

Testing your homepage and assuming the rest is fine. Your homepage might be optimized but your product pages, blog posts, or contact page might be slow. Test the pages that matter most for your business.

Only looking at the performance score. The score is a summary. The Core Web Vitals and diagnostics are what actually tell you what to fix. A score of 60 with clear image issues is more actionable than a score of 75 with no clear diagnostics.

Ignoring Field Data. If your Field Data shows failures but your Lab Data looks fine, your real users are having a bad experience. Don't dismiss it just because the simulation looks good.

What to Do After You Test

Testing is step one. Acting on the results is where the value is.

1. Identify your worst Core Web Vital. Which of LCP, CLS, or INP is failing? That's your priority.
2. Read the diagnostics. What specific issues is Lighthouse flagging? Sort by potential savings.
3. Decide what you can fix yourself and what needs help. Image compression is DIY. Render-blocking JavaScript is not. Read our guide on whether to hire someone or DIY if you're not sure.
4. Fix the issues. Start with the highest-impact items.
5. Test again. Re-run PageSpeed Insights and check if your scores improved.
6. Monitor over time. Don't just test once. Check monthly, or use Google Search Console's Core Web Vitals report to monitor real-world performance continuously.

The Fastest Way to Get Started

You can go to pagespeed.web.dev right now and test your site for free. Or you can run a free audit on WL Tech that gives you Lighthouse scores, Core Web Vitals, accessibility checks, and a prioritized list of what to fix, all in one report.

Either way, the point is to stop guessing and start measuring. You can't fix what you haven't tested.