PageSpeed Insights (PSI) is Google’s free tool for testing how fast and responsive a web page is. You enter a URL at pagespeed.web.dev and it returns two different kinds of result: real-world measurements from Chrome users who visited the page, and a simulated lab test with a 0 to 100 performance score and a list of suggested fixes.
How PageSpeed Insights works
The top section shows field data from the Chrome User Experience Report, known as CrUX. It collects anonymised timings from real Chrome users over a rolling 28-day window and reports the three Core Web Vitals:
| Metric | What it measures | Good threshold |
|---|---|---|
| Largest Contentful Paint (LCP) | How long the main content takes to appear | 2.5 seconds or less |
| Interaction to Next Paint (INP) | How quickly the page responds to taps, clicks and key presses | 200 milliseconds or less |
| Cumulative Layout Shift (CLS) | How much the layout jumps while loading | 0.1 or less |
A page passes when at least 75% of visits meet each threshold. If the page has too little traffic for its own data, PSI falls back to the whole site’s data, or shows nothing.
The lower section is lab data from Lighthouse. It loads the page once on a simulated mid-range phone with a throttled connection, then scores performance and lists opportunities such as oversized images, render-blocking scripts and unused code. It also scores accessibility, best practices and basic SEO.
Why it matters
The field data is the part Google’s ranking systems care about, because it reflects what real visitors experience. The lab data is a diagnostic. It tells you why a page might be slow and gives repeatable results you can compare before and after a change.
For a UK small business, the most common pattern is a site that feels fast on a fibre connection in the office and slow on a phone on a train. PSI’s mobile test is closer to the second experience, which is the one most of your search visitors have.
Common mistakes
- Treating the lab score as the result. A score of 45 with passing field data is a better position than a score of 90 with failing field data.
- Testing once. Lab scores vary from run to run. Test three times and look at the pattern.
- Fixing every suggestion in order. The list is not ranked by business impact. A few big causes usually explain most of the delay.
- Testing only the homepage. Service, product and blog templates often behave very differently.
- Expecting changes to show at once in field data. Because it is a 28-day window, improvements take weeks to appear in full.
How to act on it
Test one representative URL for each main template on mobile. Note whether field data passes, and which metric fails. Then use the lab diagnostics to find the cause of that specific metric: for slow LCP, look at the main image, server response time and render-blocking files; for poor INP, look at heavy scripts such as chat widgets, tag managers and sliders; for CLS, look at images without dimensions, late-loading banners and web fonts.
Fix the template rather than individual pages, then track the Core Web Vitals report in Search Console over the following month. If the causes sit in hosting, theme code or third-party tags, my technical SEO service covers diagnosing and fixing them.