What is PageSpeed Insights?

PageSpeed Insights (PSI) is Google's free web performance analysis tool, available at pagespeed.web.dev, that evaluates a URL against both lab-based Lighthouse audits and real-user field data from the Chrome User Experience Report (CrUX). First launched in 2010 as a simpler scoring tool, PSI was overhauled in 2018 (v5) to integrate Lighthouse and again in 2023 to align fully with Core Web Vitals thresholds—LCP under 2.5s, INP under 200ms, and CLS under 0.1. PSI returns a 0–100 performance score derived from Lighthouse's weighted metric formula, plus a separate section showing 28-day rolling field data from CrUX at both origin and URL granularity. The field data section is what actually influences Google Search ranking signals. We run every client URL through PSI before and after deployments—it's the fastest way to spot regressions and confirm Core Web Vitals pass/fail status.

How it works

When you enter a URL into PageSpeed Insights, two things happen in parallel:

1. Lab audit (Lighthouse) PSI runs a Lighthouse analysis on Google's servers using a simulated mobile device (Moto G Power on throttled 4G, as of early 2026). This produces the 0–100 score based on six weighted metrics:

First Contentful Paint (10%)
Speed Index (10%)
Largest Contentful Paint (25%)
Total Blocking Time (30%)
Cumulative Layout Shift (25%)

These weights shifted in Lighthouse 12 when INP replaced FID in field data, though TBT remains the lab proxy for interactivity.

2. Field data (CrUX) PSI queries the Chrome UX Report API for 28-day aggregated data from real Chrome users. If the specific URL has enough traffic, you'll see URL-level data; otherwise, it falls back to origin-level data. Field metrics include LCP, INP, CLS, FCP, TTFB, and an overall Core Web Vitals assessment.

The API behind PSI is public—you can call https://www.googleapis.com/pagespeedonline/v5/runPagespeed?url=YOUR_URL&key=YOUR_KEY to get the same JSON response programmatically. We've built CI checks around this endpoint on 50+ projects to block deploys that regress performance.

curl "https://www.googleapis.com/pagespeedonline/v5/runPagespeed?url=https://example.com&strategy=mobile&key=YOUR_API_KEY"

The JSON response includes both lighthouseResult and loadingExperience objects, letting you parse lab and field data separately.

When to use it

PSI is the right tool when you need a quick, authoritative snapshot of page performance that includes real-user data. Here's when it fits and when it doesn't:

Use PSI when:

You need to check if a page passes Core Web Vitals for search ranking purposes
You want CrUX field data without setting up the CrUX API separately
You're reporting performance to a non-technical stakeholder (the UI is clear)
You want a fast pre/post comparison after a deploy
You need the Lighthouse treemap and filmstrip for a quick diagnostic

Skip PSI when:

You need to test authenticated pages (PSI can't log in)
You want to compare performance across dozens of URLs at once—use the CrUX History API or Lighthouse CI instead
You need consistent, reproducible lab results—run Lighthouse locally with --throttling-method=devtools for less variance
You're debugging layout shifts frame-by-frame—use Chrome DevTools Performance panel

We treat PSI as the sanity check, not the investigation tool. It tells you what's wrong; DevTools and WebPageTest tell you why.

PageSpeed Insights vs alternatives

Tool	Data type	Auth pages	Bulk testing	Cost
PageSpeed Insights	Lab + Field (CrUX)	No	No (1 URL at a time)	Free
Lighthouse CLI	Lab only	Yes (with scripting)	Yes (via Lighthouse CI)	Free
WebPageTest	Lab only	Yes (scripted login)	Yes (API)	Free tier + paid
CrUX Dashboard	Field only	N/A	Yes (origin-level)	Free
Chrome DevTools	Lab only	Yes	No	Free

PSI's unique advantage is that it's the only free tool that surfaces CrUX field data alongside a Lighthouse audit in one view. Lighthouse CLI gives you more control and reproducibility. WebPageTest gives you waterfall-level detail and multi-step testing. For most of our client projects, we use PSI for the field data check and Lighthouse CI in GitHub Actions for the lab regression gate.

Real-world example

On a Next.js e-commerce project, we noticed PSI field data showing the product listing page had an LCP of 3.8s (75th percentile)—failing the 2.5s threshold. The Lighthouse lab score was 72. We identified that the hero image was served as an unoptimized PNG via a third-party CDN. After switching to next/image with AVIF format and adding fetchpriority="high", the lab LCP dropped to 1.6s and the Lighthouse score jumped to 94. Four weeks later, CrUX field data in PSI confirmed the 75th percentile LCP had fallen to 2.1s, moving the page into "good" territory. That CrUX improvement is what actually mattered for the page's search ranking signal.

Frequently asked questions about PageSpeed Insights

Is PageSpeed Insights the same as Lighthouse?

No. PageSpeed Insights runs Lighthouse as its lab testing engine, but it also includes real-user field data from the Chrome User Experience Report (CrUX), which Lighthouse alone doesn't provide. Think of PSI as a wrapper: it gives you a Lighthouse audit plus 28-day aggregated data from actual Chrome users visiting that URL or origin. The field data section is arguably more important because it reflects what Google uses for its page experience ranking signals. Lighthouse scores can vary between runs due to server conditions and network simulation differences, but CrUX field data represents stable, real-world measurements.

When did PageSpeed Insights become standard?

Google launched the original PageSpeed Insights around 2010, but it was a much simpler tool focused on rule-based recommendations. The major turning point was November 2018 when Google released PSI v5, integrating Lighthouse scoring and CrUX field data into the same interface. In 2020, Google announced Core Web Vitals and PSI became the go-to tool for checking them. In March 2024, INP officially replaced FID as a Core Web Vital, and PSI updated to reflect this. By 2026, PSI is the de facto first stop for any performance conversation involving Google search rankings.

What's the alternative to PageSpeed Insights?

The closest alternative depends on what you need. For lab-only testing with more control, Lighthouse CLI (run locally or in CI) is the standard—you can test authenticated pages and get consistent throttling. For deeper waterfall analysis and multi-step scripted tests, WebPageTest remains unmatched. For field data without the lab audit, the CrUX API or CrUX BigQuery dataset gives you direct access to real-user metrics. For continuous monitoring, tools like Calibre, SpeedCurve, or DebugBear offer dashboards tracking performance over time. We typically pair PSI for quick checks with Lighthouse CI in our GitHub Actions pipeline for automated regression prevention.

Does a perfect PageSpeed Insights score guarantee good search rankings?

No. A 100 Lighthouse score in PSI confirms strong lab performance, but Google's ranking algorithm considers hundreds of factors—content relevance, backlinks, E-E-A-T, and many others. Page experience (which includes Core Web Vitals from CrUX field data) acts as a tiebreaker, not a dominant signal. Google has said this explicitly. Also, the lab score and field data can diverge significantly: a page can score 95 in the lab but fail CrUX thresholds due to slow real-world devices or heavy ad scripts loading post-audit. Focus on passing the CrUX field data thresholds rather than chasing a perfect lab number.