Audits

What is a good PageSpeed Insights score?

A Performance score of 90 to 100 is good. What the bands mean, why the score is a lab test, why it moves between runs, and what Google actually measures.

3 min read

On this page
  1. The three score bands
  2. What the score is made of
  3. Real visitors come first
  4. Why the score changes between runs
  5. Does a good score help rankings?
  6. What to watch out for when you quote a score
  7. Check speed with RadiusPilot
  8. Sources

In PageSpeed Insights, a Performance score of 90 to 100 is good, 50 to 89 needs improvement and 0 to 49 is poor. That score comes from a simulated lab test, though: how real visitors experience the page is the Core Web Vitals assessment at the top of the report.

The three score bands

ScoreBand
90 to 100Good (green)
50 to 89Needs improvement (orange)
0 to 49Poor (red)

Phones and computers get separate results, on separate tabs. Always say which one you’re quoting.

What the score is made of

PageSpeed Insights runs Lighthouse on the address in a simulated environment, then weighs five lab measures. These are the weights Chrome’s Lighthouse documentation lists first:

Lab measureWeight
Total Blocking Time30%
Largest Contentful Paint25%
Cumulative Layout Shift25%
First Contentful Paint10%
Speed Index10%

Only these measures count. The opportunities and diagnostics further down suggest fixes, but they don’t change the score.

Real visitors come first

When Google has enough visits, the top of the report shows field data: real Chrome users over the previous 28 days. The page passes the Core Web Vitals assessment when the 75th percentile of all three measures is good. If INP has too little data, LCP and CLS decide.

  • Largest Contentful Paint (LCP): the main image or text block appears within 2.5 seconds.
  • Interaction to Next Paint (INP): the page reacts to a tap or click in 200 milliseconds or less.
  • Cumulative Layout Shift (CLS): the layout moves by 0.1 or less while it loads.

A new page, or one with few visits, may not have enough data. PageSpeed Insights then shows data for the whole site, or no real-user data at all, and the lab score is all you have. The two can disagree: the lab simulates one device and connection, while field data comes from real people on theirs.

Why the score changes between runs

Run the same page twice and the number can move. Chrome’s documentation says much of that variability doesn’t come from Lighthouse itself but from changing conditions, such as:

  • A/B tests or ads that change between loads
  • Changes in how internet traffic is routed
  • Testing on different devices
  • Browser extensions that add scripts or change requests
  • Antivirus software

It suggests treating performance as a range of scores rather than one number. Run the test a few times, note the lowest and highest, and keep the date.

Does a good score help rankings?

Google recommends good Core Web Vitals “for success with Search” and says they align with what its core ranking systems seek to reward. It also says there is no single page experience signal, that good results in its reports or third-party tools don’t guarantee top rankings, and that Search “always seeks to show the most relevant content, even if the page experience is sub-par.”

So a faster page is better for visitors and fits what Google says it seeks to reward, but a score of 100 is not a ranking.

What to watch out for when you quote a score

  • Name the tab and the source. “Mobile, lab test” or “real visitors, last 28 days”.
  • Compare like with like. The same page, the same tab, tested the same day.
  • Lead with what people notice. “On a phone, your main photo takes longer than Google’s 2.5-second target to appear” means more to an owner than a bare score, as long as the report shows it.
  • Promise no score and no ranking. Offer to measure again after the work, the same way.

Speed is one of ten checks in the website audit checklist, which also covers HTTPS, the phone layout and booking. Finding local businesses that need a new website shows where the audit fits in prospecting.

Check speed with RadiusPilot

Site speed runs Google’s PageSpeed test on any address, on a phone or a computer: the four Lighthouse scores, the lab timings, and real-visitor data when Google has it. It runs in the web app, with nothing to install. Run audit includes the same test in a full website audit: see what the website checks look at.

Sources

Checked against these pages on November 3, 2026.

Get started

Find these businesses in minutes

RadiusPilot scans a neighbourhood, checks each business’s website and listing, and puts the best prospects first.