What does the PageSpeed score actually mean?
The PageSpeed score is not your site's speed. It is a weighted average of five lab metrics from one run on a simulated mid-range phone with a throttled network. Real visitor experience shows up in field data, and the score varying between runs is normal.
The score is the most visible and most misunderstood part of PageSpeed. People chase it, argue about it, and fix things because of it that did not need fixing.
So it is worth knowing exactly what it is made of.
What the score is made of
The Lighthouse performance score is a weighted average of five metrics, and the weights are not equal. The heaviest are Largest Contentful Paint and Total Blocking Time — together more than half.
The practical consequence: fixing what weighs little barely moves the score. A couple of seconds off LCP moves it immediately.
Why the score jumps
The same page measured three times gives three different numbers — often 8 to 10 points apart. That is not a bug, it is the nature of the measurement: each run is one visit over one network at one moment.
So 92 and 86 are not a difference. 92 and 61 are.
Lab and field
Two numbers on the same page measure different things:
- lab (Lighthouse) — one run under controlled conditions. Good for diagnosis: it tells you what is slow;
- field (CrUX) — real visitor data over 28 days. Good for the verdict: it tells you whether it is actually slow.
When they disagree, the field is right. The lab simulates a mid-range phone on a throttled network; your visitors may be faster or slower than that.
A small site often has no field data at all — it needs a few hundred visits over 28 days. Then the lab plus common sense is what you have.
What to do instead of chasing it
Look at the metrics, not the score. LCP above 2.5 s and INP above 200 ms are concrete things with concrete causes.
Start with images. Almost always the single biggest win: right format, right size, dimensions declared.
Do not chase 100. The difference between 90 and 100 is usually invisible to a visitor, and the cost is disproportionate.
Related articles
The same subject from another angle — each answers its own question:
- Why is my website slow and where to start
- Core Web Vitals with no data: what it means
- Technical SEO audit: what to check and when
How this feeds into the score is described under what we check.
Sources
What this text rests on — primary sources, not retellings:
Frequently asked
Does the score affect rankings?
Not directly. Ranking uses **field** Core Web Vitals data, not the Lighthouse score. And even that is one signal among many: a fast page with no answer on it will not rise.
Why is the mobile score lower than desktop?
Because it is measured on a simulated mid-range phone with a throttled network. That is deliberately the harsher condition — most real visitors arrive on a phone.
What does "not enough data" mean?
That there were not enough real visits over 28 days to form field statistics. It is neither an error nor a penalty — just low traffic.
Is it worth getting every metric into the green?
It is worth getting none of them into the red. Green everywhere is nice, but the last few percent usually cost more than they return.
Check yours
A free scan shows which of the things described here your site already does, and which it does not.