What is Google PageSpeed Insights

What it does

PSI runs Lighthouse and surfaces CrUX field data to score performance and diagnose what to fix.

• Field data: real-user LCP, INP and CLS from the Chrome UX Report (pass/fail Core Web Vitals).
• Lab data: a Lighthouse performance score with detailed timings.
• Prioritised opportunities and diagnostics (render-blocking resources, image sizing, unused code).
• Separate mobile and desktop analysis.
• A shareable report; the same engine is available via the PageSpeed Insights API.

Where it fits in your stack

PSI is the performance-diagnosis layer for technical SEO. It overlaps with the Core Web Vitals report in Search Console, but PSI works on any single URL and adds Lighthouse-level detail on how to fix it. In dev workflows the Lighthouse engine underneath PSI runs locally and in CI.

Strengths and limits

Strength: free, authoritative on Core Web Vitals (it uses the same CrUX data Google ranks on), and specific about fixes. Limit: lab scores fluctuate between runs and can mislead if over-weighted; field data is missing on low-traffic pages; and the score is not a direct ranking multiplier — passing Core Web Vitals is the goal, not chasing a perfect 100.

How REO Rank uses it

We use PSI field data to confirm whether real users pass Core Web Vitals, then work the lab diagnostics into our technical-SEO and web-dev backlogs. We validate against the Search Console Core Web Vitals report and CrUX rather than fixating on a single Lighthouse score.

Why it matters

PageSpeed Insights (PSI) is Google's free page-performance tool, and it is important because it reports both lab and field data for a URL in one place — including the real-user Core Web Vitals Google actually uses for ranking. It matters as the go-to first check on a page's speed and page-experience health, and crucially it distinguishes the two kinds of data most people conflate: the simulated Lighthouse lab score, and the 28-day real-user field data from the CrUX dataset. Reading both correctly is the difference between chasing a vanity lab number and fixing what Google actually measures.

• Google's free tool reporting both lab and field performance data for a URL
• Surfaces the real-user Core Web Vitals Google uses for ranking
• Distinguishes the Lighthouse lab score from the CrUX field data

Lab vs field data

PSI shows two datasets and they answer different questions. The field data (labelled as real-user experience, from the Chrome User Experience Report) is the 75th-percentile of actual Chrome users over 28 days — this is what determines whether a page passes Core Web Vitals for ranking. The lab data comes from a single Lighthouse run in a controlled environment and is useful for debugging and reproducing issues, but it is a simulation, not what Google ranks on. The classic mistake is optimising to make the Lighthouse score green while the field data still fails; PSI's value is showing you both so you optimise the right one.

• Field data: real Chrome users, 75th percentile over 28 days — what Google ranks on
• Lab data: a single Lighthouse simulation, useful for debugging not ranking
• A page can pass the lab score yet fail field data (or vice versa)
• Optimise to the field data; use the lab run to diagnose

In practice

Use PSI as a first diagnostic: enter a URL, read the field data to see whether it genuinely passes Core Web Vitals, then use the Lighthouse lab diagnostics and opportunities to identify what to fix — render-blocking resources, oversized images, main-thread blocking, layout shifts. Fix the underlying issues (often via development work on the rendering path, image pipeline and JavaScript), then re-verify. Because field data lags 28 days, confirm real-world improvement in Search Console's Core Web Vitals report over time rather than expecting the field number to move instantly. PSI diagnoses; the fixes usually live in the code.

• Read the field data first to judge real Core Web Vitals pass/fail
• Use lab diagnostics to identify specific issues to fix
• Field data lags 28 days — confirm improvement over time in Search Console
• PSI diagnoses; the actual fixes are usually engineering work