PageSpeed Insights Google: What the Numbers Actually Mean
PageSpeed Insights gives you a score out of 100 and a wall of coloured labels. Most people glance at the number, feel vaguely worried, and then close the tab. That is understandable. The report is not exactly easy reading. But those numbers are telling you something specific about how your site performs for real visitors. Once you understand what each part actually measures, the report stops being overwhelming and starts being useful.
On this page
What PageSpeed Insights Actually Is
PageSpeed Insights is a free tool from Google. You paste in a URL and it analyses the page, then reports back on load performance and user experience. It pulls data from two places, a lab test run on a simulated device, and real-world data collected from actual Chrome users who have visited your site.
That distinction matters more than most people realise. The lab score is what the tool calculates right now. The field data is what real visitors have actually experienced over the past 28 days. The two can differ a lot, especially if your audience uses older phones or slower connections.
The Score: What 90, 50 and 30 Actually Represent
Google colours the score in three bands. Green (90 to 100) means the page is performing well. Orange (50 to 89) means there is room to improve. Red (0 to 49) means the page has real problems that are likely affecting visitors.
The score is a weighted average of several individual metrics. A page can score 78 overall and still be failing on the one metric that matters most for search. So the headline number is a rough guide, not the full picture. What sits beneath it is where the detail lives.
It is worth knowing that which score you focus on depends on the device tab you are reading. The mobile score and the desktop score are measured separately, and mobile nearly always comes in lower.
The Six Core Metrics and What They Measure
Six metrics feed into the overall score. Three carry the most weight.
- Largest Contentful Paint (LCP) measures how long it takes for the biggest visible element on the page to load. Usually that is a hero image or a large heading. Google considers anything under 2.5 seconds good.
- Total Blocking Time (TBT) measures how long the page is unresponsive to clicks or taps while scripts are loading. High TBT means the page looks ready but does nothing when you try to use it.
- Cumulative Layout Shift (CLS) measures how much the page jumps around as it loads. If a button moves just as you tap it, that is a CLS problem.
The other three are First Contentful Paint, Speed Index, and Time to Interactive. They contribute to the score, but fixing LCP, TBT and CLS tends to move things the most.
Field Data vs Lab Data: Which One Google Cares About
The field data section sits near the top of the report. It shows Core Web Vitals as Google has actually measured them for your visitors. These are the numbers that feed into how Google uses page experience as a ranking signal.
If your site is new or low traffic, the field data section may show ‘insufficient data’. In that case, the lab data is all you have to go on. It is still worth acting on, but understand you are working from a simulation, not a measurement of your real audience.
A common issue we see across sites is a passing lab score sitting alongside a failing field data result. That usually means the lab test ran on a fast connection with no cached assets, while real visitors on phones are getting a noticeably slower experience. The lab score can flatter you.
The Opportunities Section: What Is Worth Fixing
Below the metrics, PageSpeed Insights lists Opportunities and Diagnostics. Opportunities show specific changes that could improve load time, with an estimated saving shown in seconds. Diagnostics flag technical issues without a direct time estimate.
Focus on Opportunities first. Common ones include serving images in next-gen formats, removing unused JavaScript, and eliminating render-blocking resources. These are real, fixable things. Some take five minutes. Others take a developer a day.
Not every Opportunity is worth chasing. If an item saves 0.05 seconds and requires rewriting a plugin, leave it. Spend the time on the ones with a meaningful estimated saving. That is where the score movement actually comes from.
A Realistic View of What Good Looks Like
A score of 90 or above on mobile is genuinely hard to hit for most sites, especially those running a CMS with third-party plugins and tracking scripts. A score in the high 70s with all Core Web Vitals passing in field data is a solid result.
PageSpeed is not a one-time fix. Add a new plugin, embed a video, update a theme, and the score shifts. It needs checking periodically, not just once. The sites that hold their scores are the ones where someone is paying attention to what goes in, not just what gets built at the start.