PageSpeed by Google: What It Actually Measures and Why It Matters
Most people run PageSpeed by Google once, glance at the score, and feel either relieved or confused. The number itself tells you almost nothing useful. What matters is what sits underneath it, the specific signals Google is actually measuring, why they affect your rankings and your visitors, and which ones are worth fixing first. This is a plain-English look at what the tool does and what to take seriously.
On this page
The score is a summary, not the point
Google’s PageSpeed Insights gives you a number out of 100. It feels definitive. It isn’t. The score is a weighted average of several individual metrics, and a site scoring 72 can outrank one scoring 85 if the right signals are in better shape. Chasing the headline number is the wrong goal.
The tool runs two separate tests, one simulated on a mobile connection, one on desktop. The mobile result is almost always lower, and that’s the one Google weighs more heavily in search rankings. A site that looks fine on a laptop can still be slow where it matters most.
The six metrics it actually tests
PageSpeed Insights measures a handful of specific things. Some carry far more weight than others in the final score calculation.
- Largest Contentful Paint (LCP) , how long before the biggest visible element on the page loads. For most sites that’s the hero image or the main heading block.
- First Contentful Paint (FCP) , the time until any content appears at all. A slow FCP means the visitor stares at a blank screen.
- Total Blocking Time (TBT) , a lab-measured proxy for how long scripts block the browser from responding to clicks. Heavy Javascript is usually the cause.
- Cumulative Layout Shift (CLS) , how much the page jumps around as it loads. Images without dimensions, fonts swapping in late, ads loading after text, all of these push the score up.
- Speed Index , how quickly content becomes visible during load, not just when it technically finishes.
- Interaction to Next Paint (INP) , a newer addition that measures how quickly the page responds after a user clicks or types something.
LCP, CLS and INP are the three Google has designated as Core Web Vitals, meaning they feed directly into search ranking signals. The others matter for user experience, but these three are what Google officially tracks.
Lab data versus field data
PageSpeed Insights shows two types of data, and most people ignore the difference. Lab data is a simulated test run in controlled conditions. Field data is real-world performance collected from actual Chrome users visiting your site, pulled from the Chrome User Experience Report (CrUX).
Field data is what Google uses for ranking. If your site doesn’t have enough traffic to generate a CrUX dataset, you’ll see no field data at all, and Google falls back on lab data instead. For low-traffic sites, this means the lab result matters more than it usually would.
The practical gap between the two can be large. A page might score 90 in lab conditions and show a poor LCP in field data, because real users are on slower connections, older phones, or have browser extensions running. Always read both.
What actually slows a page down
The tool flags a list of opportunities and diagnostics beneath the score. These are the specific things worth paying attention to. Common culprits include uncompressed images, render-blocking scripts loaded in the wrong order, unused CSS, and third-party embeds like chat widgets or video players that fire requests before the main content has even painted.
WordPress sites often struggle here because themes and plugins add their own scripts and stylesheets without much regard for what’s already on the page. The more plugins, the more requests, and the heavier the page gets. It compounds quietly over time, and most site owners don’t notice until the score has already dropped. If you want to understand what a theme is actually loading, our look at what sits under a WordPress theme covers this in more detail.
Why it matters for SEO
Google confirmed page experience signals, including Core Web Vitals, as ranking factors. A site with poor LCP and high CLS is not invisible to the algorithm. It’s at a disadvantage, particularly in competitive niches where two sites are otherwise similar in quality and authority.
The effect is rarely dramatic overnight. SEO changes bed in gradually, and page speed is no different. Fix a slow LCP and you won’t jump ten positions the next morning. But across weeks and months, a technically cleaner site accumulates a small but real edge. That compounds.
There’s also the human side. A page that takes four seconds to show anything loses visitors before they’ve read a word. No ranking matters if people leave immediately.
What not to obsess over
A score of 100 is not the target. Most high-performing commercial sites sit in the 60s or 70s on mobile and rank well regardless. The honest trade-off is this, a completely stripped-back site with no images, no interactivity and no third-party tools might score 100 and convert nobody. Real sites have real features, and features cost milliseconds.
Focus on the Core Web Vitals thresholds Google actually publishes. LCP under 2.5 seconds. CLS below 0.1. INP under 200 milliseconds. Hit those and you’re in good shape. Everything above that is diminishing returns unless you’re running a high-traffic site where even small gains add up at scale.