What is PageSpeed Insights (PSI)?

What is PageSpeed Insights (PSI)?

PageSpeed Insights (PSI) is Google's free tool for analyzing how fast a web page loads and responds. Enter a URL at pagespeed.web.dev and it shows two kinds of data: field data from real Chrome users, taken from the Chrome User Experience Report, and lab data from a Lighthouse test. It also gives a Performance score from 0 to 100 and a list of suggested fixes, with separate results for mobile and desktop.

Field data and lab data

The two halves of a PSI report answer different questions, and mixing them up is the most common reading mistake.

Field data: how do real users experience this page?

Field data comes from the Chrome User Experience Report and covers the previous 28-day collection period. PSI shows the 75th percentile of each metric: Largest Contentful Paint (LCP), Interaction to Next Paint (INP) and Cumulative Layout Shift (CLS), plus First Contentful Paint and an experimental Time to First Byte. The good thresholds are LCP at or under 2,500 ms, INP at or under 200 ms and CLS at or under 0.1. A page passes the Core Web Vitals assessment when the 75th percentile of all three is good.

If a page has too few samples, PSI falls back to origin-level data, which covers every page on the site. If the origin has too little data too, PSI shows no field data at all, which is normal for low-traffic sites.

Lab data: what does a controlled test show?

Lab data comes from a Lighthouse run that simulates a mid-tier mobile device on a mobile network, or a desktop on a wired connection. It is reproducible and good for debugging, but it can miss real-world bottlenecks, because no real user is involved.

How the PSI Performance score works

The 0 to 100 Performance score is calculated from lab metrics only. In Chrome's Lighthouse 10 documentation the weights are Total Blocking Time 30%, Largest Contentful Paint 25%, Cumulative Layout Shift 25%, First Contentful Paint 10% and Speed Index 10%. A score of 90 or above is good, 50 to 89 needs improvement and below 50 is poor. Weights change between Lighthouse versions, so treat these numbers as a snapshot.

Below the score, PSI lists opportunities, which carry an estimated saving, and diagnostics, which are findings without one. Opportunities and diagnostics do not feed the score directly, but fixing them usually improves the metrics that do.

Does PSI affect SEO?

Google recommends good Core Web Vitals for success in Search, and says its page experience guidance aligns with what its core ranking systems seek to reward. The Core Web Vitals are field metrics, so the number to watch is your field data. The lab Performance score is a diagnostic aid, not a ranking number.

The PageSpeed Insights API

PSI is also available as an API. A request to https://www.googleapis.com/pagespeedonline/v5/runPagespeed?url=... returns lighthouseResult (lab data), loadingExperience (field data for the URL) and originLoadingExperience (field data for the origin). An API key is optional, but Google recommends one for frequent automated queries. Google's docs also say it plans to stop including Chrome UX Report data in this API and recommend the CrUX API or CrUX History API for field data, so build new field-data reporting on those.

Common mistakes when reading PSI

• Chasing 100. The last points cost far more effort than the first ones. Aim for good field Core Web Vitals instead.
• Trusting a single run. Lab scores move with network conditions, test devices, A/B tests, ads and even browser extensions. Chrome recommends looking at the distribution, so run it several times and compare medians.
• Ignoring field data. A good lab score with poor field data means real users have slower devices, networks or third-party scripts than the simulation assumed.
• Expecting instant results. Field data covers 28 days, so a fix takes weeks to show up fully.
• Comparing tools directly. Other speed tests use different throttling profiles, so the same page scores differently. Pick one tool and track its trend.

PSI, Lighthouse and monitoring

PSI is an on-demand check of one URL. Lighthouse itself, run from Chrome DevTools or the command line, is faster for iterating on fixes. Real user monitoring measures every visit rather than a sample. For tracking over time, LoadFocus page speed monitoring runs Lighthouse audits on a schedule (hourly, daily or weekly) from 25 cloud regions and can alert you by email, Slack or Microsoft Teams. The free Core Web Vitals checker reads the same Chrome UX Report field data for a site, with no signup. For a step-by-step approach, see how PageSpeed Insights can improve your rankings and how to monitor Core Web Vitals with alerts and reports.

FAQ: PageSpeed Insights

Why is my PSI score so low?

The usual causes are render-blocking scripts, oversized images, layout shifts from late-loading content and heavy third-party scripts such as ads, analytics and chat widgets. PSI's opportunities and diagnostics rank them, so start at the top.

Why is my mobile score lower than desktop?

Lighthouse simulates a mid-tier mobile device on a mobile network for the mobile test, and an emulated desktop on a wired connection for the desktop test. The same page does more work on a slower device and slower network, so the score drops.

Why do my PSI scores change between runs?

Lab results vary with network routing, the test device, A/B tests, changing ads and similar factors. Run the test several times and compare medians. Field data also moves as your visitor mix changes.

Why does PSI say there is no field data?

The page and its origin both have too few Chrome samples in the Chrome UX Report. Only the lab data is shown until traffic grows. Lab data still helps with debugging.

How often should I check PageSpeed Insights?

Check it when you ship a change that affects loading. For ongoing tracking, a scheduled monitor catches regressions without anyone having to remember to run a test.