PageSpeed Insights (PSI) is a free Google tool that analyzes the performance of any URL and returns a score from 0 to 100, along with real-world field data and prioritized improvement recommendations. It's the fastest way to understand how Google perceives your page's speed and which technical issues are slowing it down.
If you've ever wondered why your site loads slowly or how to improve your Core Web Vitals, this tool is your mandatory starting point.
What Is PageSpeed Insights and Where Did It Come From
PageSpeed Insights was launched by Google in 2010 as a Firebug extension. Over the years it evolved to integrate with Lighthouse, the open-source auditing engine that Google also uses inside Chrome DevTools.
Today, PSI combines two distinct data sources:
- Lab data: a controlled simulation run with Lighthouse on Google's servers. Always available and reproducible.
- Field data (CrUX): real experiences from users who visited your page in the last 28 days, sourced from the Chrome User Experience Report. Only shown if your URL has enough traffic.
The distinction matters: lab data shows your page's potential; field data shows what your users actually experience.
Which Metrics Does PageSpeed Insights Measure
PSI evaluates six core metrics, all part of the broader web performance and Core Web Vitals ecosystem:
| Metric | What it measures | "Good" threshold |
|---|---|---|
| LCP (Largest Contentful Paint) | Time to render the largest visible element | ≤ 2.5 s |
| INP (Interaction to Next Paint) | Responsiveness to user interactions | ≤ 200 ms |
| CLS (Cumulative Layout Shift) | Visual stability (how much elements move) | ≤ 0.1 |
| FCP (First Contentful Paint) | Time to display the first pixel of content | ≤ 1.8 s |
| TTFB (Time to First Byte) | Server response time | ≤ 800 ms |
| Speed Index | How quickly the screen fills up visually | ≤ 3.4 s |
LCP, INP, and CLS are the Core Web Vitals: the only ones Google uses as a ranking signal in its search algorithm.
How the 0–100 Score Is Calculated
The performance score is not a simple average. Lighthouse applies different weights to each metric and then transforms the result with a logarithmic curve so that poor values are penalized more heavily than mediocre ones.
Approximate weights in the current version of Lighthouse:
- LCP: 25%
- INP: 30%
- CLS: 15%
- FCP: 10%
- TTFB: 10%
- Speed Index: 10%
The scoring bands:
- 90–100 (green): Good. Your page is fast for most users.
- 50–89 (orange): Needs improvement. Issues affect a portion of your users.
- 0–49 (red): Poor. The experience is bad and Google penalizes it in rankings.
A green score on mobile is harder to achieve than on desktop because Lighthouse simulates a slow 4G connection and a mid-range device — the realistic scenario for most users across Latin America and beyond.
How to Read the Report Step by Step
1. Read field data first
If CrUX data is available, it reflects real user experience. A high lab score combined with red field data means the problem exists in production — likely third-party JavaScript that Lighthouse doesn't fully execute in simulation.
2. Target the highest-impact opportunities
PSI splits its diagnostics into three blocks:
- Opportunities: improvements with an estimated time saving (e.g., "Serve images in next-gen formats"). Attack these first for the best ROI.
- Diagnostics: issues without a quantifiable saving, but that still affect stability and experience (e.g., "Avoid long main-thread tasks").
- Passed audits: what's already working well. Don't fix what isn't broken.
3. Treat mobile and desktop separately
PSI runs two independent analyses. Improve mobile first — it has the greater SEO impact and is closer to your real users' experience.
4. Re-measure after every change
The score fluctuates slightly between runs due to network variation on Google's servers. Run three measurements and average them before declaring an improvement confirmed.
Common Mistakes When Interpreting PageSpeed Insights
Many site owners fall into the same traps:
- Obsessing over the round number. A site scoring 89 with good field Core Web Vitals will rank equally well or better than one scoring 92 with high CLS in the field.
- Ignoring field data when it exists. Field data is what Google actually measures for ranking. Lab data is a diagnostic guide, not the end goal.
- Measuring only the homepage. Google measures each URL individually. A slow product landing page hurts rankings for that specific page.
- Skipping mobile. Google's index is mobile-first. The desktop score is irrelevant to ranking unless compared with mobile.
How to Improve Your Score: First Steps
If your score is in the red or orange zone, these are the most effective starting points:
- Images: convert to WebP or AVIF, add explicit
widthandheightattributes, useloading="lazy"on images below the fold. - Cache: set proper
Cache-Controlheaders for static assets (minimum 1 year for hashed assets). - Server: if TTFB exceeds 800 ms, the problem lies in your hosting. A well-tuned hosting plan with PHP-FPM and OPcache enabled makes a measurable difference.
- JavaScript: defer or eliminate third-party scripts that block rendering. Every analytics tracker, chat widget, or embed has a cost.
- Critical CSS: inline only the CSS needed for the first render and defer the rest.
If you need help implementing these optimizations or aren't sure where to start, the team at elenlace.com offers web performance audits with a clear diagnosis and action plan.
Key Takeaways
- PageSpeed Insights combines lab data (Lighthouse) and real field data (CrUX) in a single report.
- The 0–100 score weights six metrics; the three Core Web Vitals (LCP, INP, CLS) carry the most weight and are Google's ranking signal.
- Green = 90–100, orange = 50–89, red = 0–49. Aim for green on mobile.
- Field data takes priority over lab data for understanding real user impact and SEO effect.
- Opportunities with the highest estimated time savings are the most efficient place to start improving.
- Always measure on mobile and across your key pages, not just the homepage.
Ready to push your score into the green? Reach out at elenlace.com and we'll review your site together at no cost.
FAQ
Does PageSpeed Insights directly affect SEO?
Yes, but through the Core Web Vitals (LCP, INP, CLS) measured in field data, not through the numeric score itself. Google uses field data as a ranking signal; improving your lab score without an impact on field data won't move the SEO needle.
How often should I measure my site?
Measure after every significant deployment and at least once a month in production. Field data updates every 28 days, so recent changes can take up to a month to appear in CrUX.
Why does my score vary between measurements?
Lighthouse runs from Google's servers under variable network conditions. A swing of ±5 points is normal. That's why averaging several runs before drawing conclusions is the recommended approach.
Is PageSpeed Insights the same as Lighthouse?
Not exactly. Lighthouse is the open-source auditing engine; PageSpeed Insights is Google's web interface that uses it internally and adds CrUX field data on top. You can also run Lighthouse directly inside Chrome DevTools or via the command line.
Further reading
Other providers and guides worth comparing:
