PageSpeed Insights

PageSpeed Insights (PSI) is Google's free tool for measuring web page performance, and provides actionable insights into improving the performance and UX of your website.

Its analysis of a page returns metrics and specific improvement recommendations for mobile and desktop. It's powered by Lighthouse (Google's open-source auditing tool) and the Chrome User Experience Report (CrUX), which contains real-world performance data from actual Chrome users.

Data sources, and important notes

Lab data (Lighthouse)

PSI runs Lighthouse on the URL in a controlled environment — simulated Moto G4 mobile device, simulated 4G connection. The results are reproducible and useful for diagnosing issues, but they're a single snapshot, not real-world experience.

Field data (CrUX)

Real performance data collected from millions of Chrome users over the previous 28 days. This is the data Google uses for the Page Experience ranking signal. CrUX field data* is the source of truth for whether a page actually performs well in practice.

(*Field data is only available for URLs with enough Chrome traffic; lower-traffic pages show only lab data.)

Core Web Vitals

Three metrics make up the Core Web Vitals of PageSpeed Insights.

Largest Contentful Paint (LCP)

How long it takes for the largest visible content element (usually a hero image or headline) to render.

• Good: ≤ 2.5 seconds
• Needs improvement: ≤ 4.0 seconds
• Poor: > 4.0 seconds

Most LCP issues are server response time, render-blocking resources, or oversized hero images.

Interaction to Next Paint (INP)

How long it takes for the page to visually respond after a user interaction (click, tap, key press). Replaced First Input Delay (FID) as a Core Web Vital in March 2024.

• Good: ≤ 200 milliseconds
• Needs improvement: ≤ 500 milliseconds
• Poor: > 500 milliseconds

INP issues are usually heavy JavaScript that blocks the main thread.

Cumulative Layout Shift (CLS)

How much the page layout shifts unexpectedly during loading. Expressed as a unitless score.

• Good: ≤ 0.1
• Needs improvement: ≤ 0.25
• Poor: > 0.25

CLS is caused by images and ads loading without reserved space, late-injected fonts, and dynamic content insertion; the impact of this is that the page layout continually changes as additional information is received on the size and appearance of objects that have already been loaded, creating a jarring experience.

Other metrics

• First Contentful Paint (FCP) — when any content first renders.
• Time to First Byte (TTFB) — server response time.
• Speed Index — how quickly content visually populates.
• Total Blocking Time (TBT) — proxy for INP in lab tests.

The score

The Lighthouse Performance score (0–100) is a weighted average of lab metrics. The current weighting (Lighthouse 10+):

• LCP: 25%
• TBT: 30%
• CLS: 25%
• FCP: 10%
• Speed Index: 10%

Score bands:

• 90–100: Good
• 50–89: Needs improvement
• 0–49: Poor

It's important to note that the PSI score is an indicator, not a goal. The actual ranking signal is the Core Web Vitals field data, not the lab score. A page can score 100 in lab and still fail field data, and vice versa.

PSI opportunities and diagnostics

Common opportunities

• Eliminate render-blocking resources — defer or async non-critical JS/CSS.
• Properly size images — don't serve a 4000px image into an 800px slot.
• Serve images in next-gen formats — WebP, AVIF.
• Defer offscreen images — use lazy loading.
• Reduce unused JavaScript — split bundles, tree-shake unused code.
• Reduce unused CSS — remove or defer.
• Minify JavaScript and CSS.
• Enable text compression — Gzip or Brotli.
• Preconnect to required origins — for third-party domains.
• Avoid enormous network payloads — total page weight.

Common diagnostics

• Avoid long main-thread tasks.
• Reduce JavaScript execution time.
• Avoid excessive DOM size.
• Ensure text remains visible during webfont load.
• Image elements have explicit width and height.
• Reduce third-party impact.

Interpreting results

Lab vs. field divergence

• Lab good, field poor — means the controlled test doesn't match real users. Causes: real-world device fragmentation, slower networks, third-party scripts behaving differently in production, geographic distance from CDN.
• Field good, lab poor — usually because users are on faster connections than the simulated test environment. Less of a concern; field data is what matters for ranking.

Score fluctuation

Lab tests are run in different conditions each time. A swing of half a dozen points between runs is normal and doesn't automatically indicate a problem.

Mobile vs. desktop

Mobile is throttled more aggressively in the lab test and is the primary scoring path Google uses. Desktop scores are regularly 20+ points higher and shouldn't be the primary metric.

SEO impact

Core Web Vitals are confirmed ranking factors as part of the Page Experience signal. The impact is modest, but real. It does not make up for poor content, but when competing among otherwise-similar pages, it can be a tie-breaker.

The bigger impacts are indirect:

• Faster pages have lower bounce rates and higher engagement, both of which feed into broader ranking signals.
• Mobile-first indexing means a slow mobile experience disadvantages the page even if the desktop score is high.
• Slow pages convert worse, regardless of ranking.

Related tools

• Search Console → Core Web Vitals — site-wide field data, grouped by URL pattern.
• Chrome DevTools → Performance and Lighthouse panels — local Lighthouse runs with full traces for debugging.
• WebPageTest (webpagetest.org) — more granular control over test conditions.

Common misconceptions

• "I need a 100 score." No, you don't. In reality, a score of 100 is a technical impossibility for the majority of sites. "Good" thresholds on each Core Web Vital are the real goals.
• "The score is all that matters." A page can score 90 and still feel slow and convert poorly if the metrics that hurt user experience aren't addressed.
• "Fixing PSI will rocket my rankings." Page Experience is one signal among hundreds. As touched on above, performance gains alone won't move poor-content pages to the top of search results.

The right framing: PSI is a diagnostic tool. Use it to identify and prioritise fixes that improve real user experience and Core Web Vitals field data.

Performance

Disclaimer: All information contained herein is for informational purposes only. It is not advice or instructional.