Alejandro Rioja.
SEO

How To Use Google Page Speed Insight?

Alejandro Rioja
Alejandro Rioja
7 min read
TL;DR

Google PageSpeed Insights combines real-user field data (CrUX) with Lighthouse lab data to score your site on Core Web Vitals — LCP, INP, and CLS. Use it to diagnose slow pages and prioritize fixes that actually move rankings.

Free newsletter

Every Wednesday. 28,400+ operators. Zero fluff.

Table of contents

Open Table of contents

What is page speed?

Page speed is how fast a browser delivers usable content to a visitor. There are two common interpretations: the time to first meaningful content above the fold, and the full page load time.

Factors outside your control include the visitor’s connection speed and device. What you can control: image size and format, JavaScript execution, server response time, and caching.

Why does page speed matter?

Google uses page speed — specifically Core Web Vitals — as a ranking signal. Slow pages hurt both rankings and conversion. The relationship is direct: as load time increases, bounce rate climbs. A page that takes five seconds to load loses a meaningful portion of visitors before they ever see your content.

Beyond rankings, speed matters for revenue. Every second of delay reduces the probability that a visitor completes a goal.

What is Google PageSpeed Insights?

PageSpeed Insights is a free Google tool that analyzes any public URL and returns two types of data:

  1. Field data — real-user measurements from the Chrome User Experience Report (CrUX), showing how actual visitors experienced the page over the past 28 days.
  2. Lab data — a simulated Lighthouse audit run against the page, which gives reproducible numbers for diagnosing issues.

The tool gives a score from 0–100. Green (90–100) is good, yellow (50–89) needs work, red (0–49) is poor.

How to use PageSpeed Insights

Go to pagespeed.web.dev, paste your URL, and click Analyze. The report has four sections worth understanding.

Field Data — Core Web Vitals (CrUX)

This is the most important section for SEO. It shows how real users experienced your page. As of 2026, Core Web Vitals are:

  1. Largest Contentful Paint (LCP) — time for the largest visible content element (usually a hero image or headline) to load. Target: under 2.5 seconds.
  2. Interaction to Next Paint (INP) — measures responsiveness across all user interactions throughout the page session. INP replaced First Input Delay (FID) in March 2024. Target: under 200ms.
  3. Cumulative Layout Shift (CLS) — measures visual stability; how much the layout jumps during load. Target: under 0.1.

First Contentful Paint (FCP) also appears as a supplementary metric but is not a Core Web Vital.

If your site has insufficient traffic, CrUX field data may not be available — in that case, only lab data shows.

Lab Data — Lighthouse

Lab data runs a controlled Lighthouse simulation. Metrics include:

  1. First Contentful Paint (FCP) — time for the first content to appear.
  2. Speed Index — how quickly content visually fills in during load.
  3. Total Blocking Time (TBT) — total time main-thread tasks blocked interactivity; a lab proxy for INP.
  4. LCP — same metric as field data, measured in a controlled environment.
  5. CLS — same metric as field data.

Lab data is useful for debugging because it’s reproducible and doesn’t require real traffic.

Opportunities and Diagnostics

This section tells you what to fix. Opportunities show specific changes with estimated savings (e.g., “Serve images in next-gen formats — potential savings: 1.2s”). Diagnostics flag issues without a direct time estimate, like render-blocking resources or excessive DOM size.

Prioritize opportunities by their estimated time savings and feasibility.

Passed Audits

These are checks your page already passes. Useful for confirming that previous work held.

How to improve your PSI score

These are the high-leverage fixes I’ve used on real sites.

1. Compress and convert images

Large images are the most common culprit for poor LCP. Convert to WebP or AVIF — both compress significantly better than JPEG at equivalent quality. Most modern image optimization tools and CDNs do this automatically.

For WordPress sites, Smush handles bulk compression. For standalone compression, Squoosh (Google’s own tool) is reliable and free. Set explicit width and height attributes on images to prevent layout shift (CLS).

2. Implement browser caching

Browser caching stores static assets locally so returning visitors don’t re-download them. Set appropriate cache headers on your server or CDN.

For WordPress, W3 Total Cache remains a solid option. If you’re on a modern hosting stack with a CDN, the CDN often handles this automatically.

3. Minify HTML, CSS, and JavaScript

Removing whitespace, comments, and redundant code reduces file sizes. Most build tools (Vite, webpack, esbuild) do this automatically in production. For WordPress, caching plugins like W3 Total Cache include minification. PSI will flag this under Opportunities if it’s a meaningful issue.

4. Reduce JavaScript execution time

INP problems usually trace back to heavy JavaScript blocking the main thread. Audit your third-party scripts — analytics, chat widgets, ad scripts — and defer or lazy-load anything not critical to initial render. This is the hardest fix but often the most impactful for INP.

5. Use a CDN

A content delivery network serves assets from edge nodes geographically close to visitors, reducing latency. For most sites, this is one of the easiest wins. Cloudflare’s free tier works well for static assets.

A note on AMP

Google’s Accelerated Mobile Pages (AMP) was a project to enforce fast mobile pages through a restricted HTML format. By 2026, AMP is no longer a ranking signal and Google has largely deprioritized it. Focus on Core Web Vitals instead — a well-optimized standard page will outperform a mediocre AMP page for both users and rankings.

Google PageSpeed Insights — 2026 FAQ

What happened to First Input Delay (FID)?

FID was retired as a Core Web Vital in March 2024 and replaced by Interaction to Next Paint (INP). FID only measured the delay on the very first user interaction. INP measures responsiveness across all interactions throughout the page session, which is a more complete picture of how a page feels in use. If you have old posts or audits referencing FID as a Core Web Vital, they’re outdated.

What’s the difference between field data and lab data in PSI?

Field data comes from real Chrome users visiting your page over the past 28 days, aggregated in the Chrome User Experience Report (CrUX). Lab data is a synthetic Lighthouse audit — a simulated load in a controlled environment. They often differ. Field data reflects real-world conditions (varied devices, connections, caching); lab data is reproducible and better for debugging. Both matter: field data determines whether you pass Core Web Vitals; lab data tells you what to fix.

Does a PSI score of 100 guarantee good rankings?

No. PSI score is a diagnostic tool, not a ranking signal itself. What matters for rankings is whether your Core Web Vitals field data passes the thresholds: LCP under 2.5s, INP under 200ms, CLS under 0.1. A page can score 95 in the lab but fail field data due to real-world conditions. Focus on the field data section, not the lab score.

How often should I run PSI audits?

Run PSI after any significant change to your site (new page template, new third-party script, major content update). For high-traffic pages, set up ongoing monitoring via Google Search Console, which surfaces Core Web Vitals issues at scale without requiring manual PSI checks for every URL.

Related reading: What is SEO? · Top 10 Editorial Tools To Keep Your Content Organized · 20 Best SEO Tools

The shorter version

If you’re reading this because the workflow it describes is eating your week, that’s the kind of loop I build AI agents for. Two build slots open at a time.

Updated for May 2026

Google’s 2026 story is AI Overviews everywhere: the SGE experiment from 2023 graduated to a default feature in May 2024 and now appears on an estimated ~60% of US informational queries. For SEO and ad operators:

The “how Google makes money” answer in 2026: still Search ads (dominant), but YouTube ads, Cloud, and Subscriptions (YouTube Premium + Google One) are all material lines now.

Keep reading

Get the AI playbook in your inbox

Every Wednesday. 28,400+ operators. Zero fluff.

⌕
↵ ↵ to see all results esc esc to close