LilyTech

What Is PageSpeed Insights? How to Check Website Speed

Đoàn Lê Minh Anh

What is PageSpeed Insights? How to check website speed and assess website performance

What is PageSpeed Insights? How to check website speed? PageSpeed Insights is a free Google tool that helps website owners assess page performance on mobile devices and desktops. Along with an overall score, it provides page-loading experience metrics, real-user data, and technical recommendations for improvement.

Understanding PageSpeed Insights reports helps businesses avoid chasing a single number. Instead, you can identify issues that affect users, prioritize appropriate solutions, and track results after making improvements.

What is PageSpeed Insights and how does it work?

PageSpeed Insights, often abbreviated as PSI, is a website performance analysis tool provided by Google. Enter the URL of a specific page, run the test, and get a report on its performance on mobile or desktop. Each run analyzes the requested page at that moment, so results can vary depending on the content, server, network, and testing conditions.

Specialist analyzing website performance with PageSpeed Insights at a Vietnamese business office
PageSpeed Insights combines test data and real-user experience to help businesses assess website performance.

The report combines two different types of data. Lab data is generated in a controlled testing environment, primarily using Lighthouse. Field data is collected from the experiences of eligible Chrome users who visited the page over a recent period, typically based on 28 days of CrUX data. Not every page has field data: new pages or pages with limited qualifying traffic may not show this section.

These two data sources complement rather than replace each other. Lab data is useful for quickly checking a change or identifying technical opportunities; field data shows how people experience the page in real-world conditions. Review both before deciding what to optimize.

Which metrics make up Core Web Vitals?

Core Web Vitals are a group of page experience metrics Google uses to assess important aspects such as how quickly content appears, responsiveness, and layout stability. The thresholds below are evaluated at the 75th percentile of visits, separately for mobile and desktop:

Team of specialists reviewing Core Web Vitals in a PageSpeed Insights report at the office
LCP, INP, and CLS reflect a website’s loading speed, responsiveness, and layout stability.
  • LCP (Largest Contentful Paint): measures how long it takes for the largest piece of content in the initial viewport to appear. A good result is 2.5 seconds or less. A high LCP is often associated with slow server responses, large images, or critical resources that load late.
  • INP (Interaction to Next Paint): measures how responsive a page is to user interactions, such as tapping a button or selecting an item. A good result is 200 milliseconds or less. Poor INP can be caused by heavy JavaScript, long-running tasks, or inefficient interface processing.
  • CLS (Cumulative Layout Shift): measures unexpected movement of page elements while a page loads. A good result is 0.1 or less. Failing to specify fixed dimensions for images or ads, or adding content late, can cause the layout to shift.

To achieve good Core Web Vitals, the metrics must meet their respective thresholds in user data rather than in just one test run. Depending on the report, PSI also displays additional metrics such as First Contentful Paint or Time to First Byte. These provide diagnostic context but should not be confused with the three current Core Web Vitals.

How to check and read a PageSpeed Insights report

  1. Enter the correct URL: Test the page you want to optimize, not just the homepage. Category pages, product pages, articles, and landing pages may use different components, images, and code.
  2. Review each device separately: Start with mobile if most of your customers visit on their phones, then check desktop. Results can differ significantly between devices.
  3. Review field data first: If field data is available, check whether the page or group of pages is rated good, needs improvement, or poor for each Core Web Vital. Note the scope of the data: some reports cover a specific URL, while others represent a group of similarly structured URLs.
  4. Compare with lab data: Review the performance score and test metrics to identify issues that can be reproduced. Scores are generally categorized as good, needs improvement, or poor, but they are not the only measure of website quality.
  5. Open the opportunities and diagnostics sections: Review recommendations such as reducing render-blocking resources, optimizing images, or removing unused code. Check the estimated time savings and the specific cause before implementing a change.
  6. Measure again after making changes: Record the URL, device, date, and items changed. Run the test several times to check for trends, and continue monitoring field data, which takes time to reflect changes.

PageSpeed Insights does not rank websites using a single overall score. A high Lighthouse score does not guarantee a good experience for every user, and a low score in one test does not mean a page will necessarily lose rankings. Use PSI alongside analytics, customer feedback, and hands-on testing on commonly used devices.

Specialist testing a website on desktop and mobile with PageSpeed Insights
Testing the correct URL on both mobile and desktop makes your PageSpeed Insights report more representative of each page’s user experience.

How to improve website speed in order of priority

Don’t try to address every warning at once. Prioritize issues that affect many users, appear on important pages, and relate to metrics that are performing poorly. For businesses, a planned optimization process is usually more effective than installing more plugins or changing the interface without measuring the results.

Technical team improving website speed based on PageSpeed Insights metrics
Addressing root causes that affect many users and poor-performing metrics is a more effective way to optimize a website.

Improve LCP and content loading speed

  • Reduce server response times by optimizing the database, configuring caching, and choosing infrastructure suited to your traffic volume.
  • Compress images, use modern formats where appropriate, and serve images at dimensions suited to their display area. Avoid loading oversized images and shrinking them with CSS.
  • Prioritize loading the main image or element in the initial viewport; don’t lazy-load the content that needs to appear first.
  • Optimize critical CSS and fonts, and reduce resources that block the browser from rendering content.

Improve INP and CLS

  • Reduce unnecessary JavaScript, break up long tasks, and defer third-party code if its functionality isn’t needed immediately.
  • Set the width, height, or aspect ratio for images, videos, and ad spaces so the page reserves room before the content appears.
  • Avoid inserting banners or notifications that suddenly push the content someone is reading down the page. If they’re needed, use a fixed area or a less disruptive way of displaying them.
  • Test menus, forms, shopping carts, and other important interactions on a real phone, rather than relying only on desktop simulations.

When advising on and implementing websites, LilyTech can help businesses review page structure, code, content, and infrastructure to identify root causes. An improvement plan should balance performance with conversions, business functionality, and operational resources, rather than focusing only on the score in a report.

Common misconceptions about using PSI

Business owner discussing how to read PageSpeed Insights reports and avoid misconceptions about scores
Don’t treat the overall score as the only conclusion; consider individual metrics, real-world data, and business goals.
  • Equating the score with user experience: The performance score is an aggregated test result, while Core Web Vitals and real-world data reflect different aspects of the experience. Review the individual metrics instead of focusing only on green or red indicators.
  • Expecting field data to update immediately: Real-user data is collected over a period of time, so newly implemented changes may not appear right away. Lab data is better suited to quickly checking a fix.
  • Treating every recommendation as mandatory: Some opportunities may have little impact on customers or may not suit the page’s functionality. Assess the benefits, risks, and costs before making changes.
  • Testing only the homepage: Customers often arrive from search engines and ads on service pages, product pages, or articles. These URLs need to be tested individually.
  • Making changes without monitoring results: Optimizing one component can affect another feature. Save results from before and after, test in a staging environment, and only then update the live website.

Conclusion

PageSpeed Insights helps businesses understand website performance through test data and real-user experience. By correctly interpreting Core Web Vitals, identifying which pages to prioritize, and making measured improvements, you can enhance the customer experience without blindly chasing scores.

If you need to assess website speed, optimize Core Web Vitals, or develop a performance improvement plan tied to business goals, contact LilyTech for advice on a solution suited to your business’s current situation and resources.

About the author

Đoàn Lê Minh Anh

Đoàn Lê Minh Anh là tác giả có nền tảng học tập tại Học viện Ngân hàng, với định hướng phát triển trong lĩnh vực nội dung và viết SEO. Minh Anh chú trọng xây dựng những bài viết rõ ràng, chính xác, dễ hiểu, cung cấp thông tin hữu ích và đáp ứng tốt nhu cầu tìm kiếm của độc giả.

Chia sẻ kiến thứcLiên hệ Zalo OA