PageSpeed Insights: How to Understand and Improve Your Website’s Performance

A fast website is easier to use. Pages that load quickly and respond smoothly help visitors find what they need, whether they are browsing on a phone, reading an article or completing a purchase. PageSpeed Insights is a free Google tool that analyses a web page and offers suggestions for improving its performance.

This guide explains what PageSpeed Insights measures, how to read its reports and which changes can make a meaningful difference.

What is PageSpeed Insights?

PageSpeed Insights (PSI) analyses a page on mobile and desktop devices. It combines two types of information:

• Field data: Real-world performance data from visitors using Chrome, when enough eligible data is available. This is drawn from the Chrome User Experience Report.
• Lab data: A controlled test run by Lighthouse. It helps identify potential performance issues, but may not reflect every visitor’s device, connection or location.

Field data shows how a page has performed for real users over time. Lab data is useful for diagnosing problems and checking the effect of changes. The two can differ because they are collected in different ways.

Understanding the PageSpeed Insights score

The Performance score is a rating from 0 to 100, based on several Lighthouse metrics. In general, a higher score indicates better performance in the test environment. However, the score is not a complete measure of user experience: it can vary between tests and should be considered alongside the underlying metrics.

Google groups Lighthouse scores into these ranges:

• 90–100: Good
• 50–89: Needs improvement
• 0–49: Poor

A score of 100 is not necessary for every website. Focus first on issues that affect visitors and on improvements that are practical for your site.

The Core Web Vitals

Core Web Vitals are a set of metrics that assess important aspects of page experience. The current set includes:

• Largest Contentful Paint (LCP): Measures how long it takes for the largest visible image or text block to appear. A good result is 2.5 seconds or less.
• Interaction to Next Paint (INP): Measures how quickly a page responds visually to user interactions, such as tapping a button or selecting a menu. A good result is 200 milliseconds or less.
• Cumulative Layout Shift (CLS): Measures unexpected movement of page content. A good result is a score of 0.1 or less.

These thresholds are assessed at the 75th percentile of visits, separately for mobile and desktop. In other words, a site should aim to meet the “good” threshold for at least 75 per cent of visits.

Other metrics in the report

PageSpeed Insights may also show Lighthouse metrics that help explain what is happening during a test. For example, First Contentful Paint (FCP) records when the first content appears, while Total Blocking Time (TBT) estimates how much the main thread is blocked during loading. These can be useful diagnostic clues, but they are not substitutes for real-world Core Web Vitals.

The report also includes opportunities and diagnostics. These may point to oversized images, render-blocking resources, unused code or long-running JavaScript. Treat them as leads to investigate rather than a checklist that must be followed blindly.

How to improve page speed

Optimise images

Images are often among the largest files on a page. Resize them to suit their display dimensions, compress them and use modern formats such as WebP or AVIF where appropriate. Specify image width and height to help prevent layout shifts. Consider loading below-the-fold images only when they are close to entering the viewport, but avoid lazy-loading the main image at the top of the page.

Reduce JavaScript work

Large or unnecessary JavaScript files can delay rendering and make a page feel unresponsive. Remove code that is not needed, split large bundles where suitable and defer non-essential scripts. Review third-party tools such as analytics, advertising and embedded widgets: they can add significant work even when your own code is lean.

Improve server response times

Slow hosting or inefficient application code can delay everything that follows. Check your server response time, use caching where suitable and consider a content delivery network (CDN) if your visitors are spread across different regions. A CDN can serve static files from locations closer to visitors.

Prevent layout shifts

Reserve space for images, video players, adverts and other elements that load after the page. Avoid inserting new content above existing content unless it follows a user action. These steps can help reduce unexpected movement and improve CLS.

Load resources thoughtfully

Fonts, stylesheets and scripts can all affect when a page becomes usable. Keep the number of fonts and font weights under control, serve files efficiently and prioritise resources needed for the visible part of the page. Preloading can help in specific cases, but excessive preloading may compete for bandwidth.

How to use PageSpeed Insights effectively

1. Test important pages. Check key templates such as the home page, product pages, articles and checkout, rather than relying on one URL.
2. Look at mobile and desktop separately. Mobile users may have slower connections and less powerful devices.
3. Start with field data. Where available, it reflects how visitors have experienced the page over time.
4. Use lab results to investigate. Review opportunities and diagnostics to identify possible causes.
5. Make one set of changes at a time. Retest and compare results so you can see whether a change helped.
6. Monitor over time. Performance can change as content, software and third-party services change.

Individual Lighthouse tests can vary because conditions such as network simulation and server load are not identical every time. If a result looks unusual, run the test again and look for patterns rather than relying on a single score.

Does PageSpeed Insights affect Google rankings?

Page experience, including Core Web Vitals, can be considered among Google’s ranking signals. However, passing every threshold or achieving a perfect PageSpeed Insights score does not guarantee higher rankings. Relevant, useful content and a good overall experience remain important. Performance work is best treated as one part of a broader effort to make a website helpful and easy to use.

In summary

PageSpeed Insights is a useful starting point for understanding website performance. Its real-user field data shows how pages perform in practice, while its Lighthouse tests help identify potential improvements. Focus on meaningful measures such as LCP, INP and CLS, investigate the causes behind poor results, and prioritise changes that make pages faster and more stable for visitors.

 

Enhancing Website Efficiency: Six Advantages of Using Page Speed Insights

1. Free and easy to use
2. Tests mobile and desktop performance
3. Combines real-user and lab data
4. Highlights Core Web Vitals
5. Suggests practical improvements
6. Helps track performance issues

 

Understanding the Limitations of Page Speed Insights: 5 Key Considerations

1. Scores can vary between tests.
2. Lab results may not reflect real users.
3. Some recommendations need technical skills.
4. Third-party scripts can be hard to optimise.
5. A high score does not guarantee better rankings.

Free and easy to use

One of the biggest advantages of PageSpeed Insights is that it’s free and easy to use. Simply enter a web page’s URL to receive a clear report on its performance on mobile and desktop, along with practical suggestions for improvement. There’s no need to install software or have specialist technical knowledge to get started, making it a useful tool for website owners, developers and beginners alike.

Tests mobile and desktop performance

PageSpeed Insights tests mobile and desktop performance separately, making it easier to see how a page behaves across different devices. This is useful because mobile visitors may have slower connections or less powerful hardware, while desktop users may experience different loading and interaction issues. Comparing both reports helps you prioritise improvements and create a smoother experience for everyone.

Combines real-user and lab data

One of PageSpeed Insights’ key strengths is that it combines real-user data with controlled lab tests. Field data shows how visitors have experienced a page across different devices and network conditions, while lab data helps identify potential performance issues in a repeatable test environment. Together, they offer both a practical view of user experience and useful clues for deciding what to improve.

Highlights Core Web Vitals

PageSpeed Insights highlights your website’s Core Web Vitals, helping you understand how visitors experience its loading speed, responsiveness and visual stability. It reports metrics such as Largest Contentful Paint (LCP), Interaction to Next Paint (INP) and Cumulative Layout Shift (CLS), and shows where performance may need attention. This makes it easier to identify issues that could frustrate users and prioritise improvements that create a smoother browsing experience.

Suggests practical improvements

One of the key benefits of PageSpeed Insights is that it suggests practical ways to improve a page’s performance. Rather than simply providing a score, its report highlights potential issues—such as oversized images, render-blocking files or unused code—and explains where changes could help. These suggestions give website owners a useful starting point for prioritising fixes and making pages faster and smoother for visitors.

Helps track performance issues

PageSpeed Insights helps you track performance issues by highlighting slow-loading elements, delayed interactions and unexpected layout shifts. Its reports make it easier to spot recurring problems, understand what may be causing them and check whether changes have improved the experience for visitors on mobile and desktop.

Scores can vary between tests.

One drawback of PageSpeed Insights is that its scores can vary between tests. Changes in network conditions, server load, device emulation and other factors may produce different results, even when the page itself has not changed. This can make small improvements difficult to measure. For a more reliable picture, run several tests under similar conditions and compare the trends rather than relying on a single score.

Lab results may not reflect real users.

One limitation of PageSpeed Insights is that its lab results come from a simulated test, so they may not reflect the experience of real visitors. The test uses specific device and network conditions, while your audience may use a wide range of phones, computers, connections and browsers. Treat lab results as useful clues for diagnosing potential issues, and compare them with field data, which is based on real users’ experiences when available.

Some recommendations need technical skills.

Some PageSpeed Insights recommendations require technical knowledge to implement safely. Changes to JavaScript, server settings, caching or how resources load may involve editing code or configuring hosting, and a quick fix can sometimes cause display or functionality problems. If you are not confident making these changes, ask a developer for help and test updates before applying them to your live website.

Third-party scripts can be hard to optimise.

Third-party scripts, such as analytics tools, advertising tags and social media widgets, can be difficult to optimise because they are controlled by external providers. They may add extra network requests, slow page rendering or make interactions less responsive, yet removing or changing them may affect important site features or tracking. PageSpeed Insights can flag their impact, but resolving it often requires balancing performance with business needs and working within the provider’s options.

A high score does not guarantee better rankings.

A high PageSpeed Insights score does not guarantee better search rankings. Page speed and Core Web Vitals are only part of Google’s broader assessment, alongside factors such as content relevance, quality and usefulness. A strong score is valuable because it can support a smoother experience for visitors, but it should not be treated as a shortcut to the top of search results. Focus on improving real-world performance while continuing to create content that meets your audience’s needs.