What is a good PageSpeed score?
Short answerPageSpeed Insights rates 90 to 100 as good, 50 to 89 as needing improvement and below 50 as poor. That score comes from one simulated Lighthouse test. What Google actually uses is the field data above it: your real visitors’ Core Web Vitals. Passing those matters more than hitting 100.
Confused by your PageSpeed report?
Send me the URL and I’ll tell you which items actually matter and which you can ignore.
The longer answer
Two different numbers on one report
PageSpeed Insights shows two things. At the top, if your site has enough traffic, is field data from real Chrome users over the past 28 days, with a pass or fail for Core Web Vitals. Below that is the coloured 0 to 100 score from a single Lighthouse lab test.
The lab score is a diagnostic. It’s run on a throttled, simulated mid-range phone and can vary from run to run. The field data is what reflects real experience, and it’s what feeds into Google’s page experience signals.
What score to aim for
Lighthouse treats 90 and above as good. On mobile, that’s a solid target for a brochure or service site. A lean, well-built small business site can usually get there.
But don’t chase 100 for its own sake. Going from 92 to 100 often means removing things that genuinely help visitors, like a booking widget or a map. A site scoring 85 that passes Core Web Vitals in the field is in a better place than one scoring 98 in the lab with poor real-world numbers.
Why your score keeps changing
Run the test three times and you may get three numbers. Network conditions, server response time and third-party scripts all vary. Ads, chat widgets and analytics tags are frequent culprits, because they load from other servers you don’t control.
Look at the trend and at the specific opportunities listed below the score: oversized images, render-blocking scripts, unused JavaScript. Those tell you what to fix. The headline number just tells you whether to bother.
What I’d do
Nick Lovett, founder of Tavro
I aim for 90 or higher on mobile for every page I build, and I usually get there because the pages are prerendered and the scripts are kept lean. But the number I report to clients is the Core Web Vitals pass in Search Console, because that’s real visitors.
When someone sends me a score in the 30s, I look at the opportunities list. It’s usually images and third-party scripts, and those are fixable without redesigning anything.
Follow-up questions
The mobile test simulates a slower phone and network. That’s closer to how many of your visitors actually experience the site.
Not directly. Google uses real-world Core Web Vitals, not the lab score. A low score is still a useful warning that real visitors may be waiting.
Your site doesn’t have enough Chrome traffic for Google to report on. Use the lab score as your guide until it does.
Keep reading
- nick@tavro.ca
- 587-439-9994
- Calgary, Alberta