By Mehdi HassanDoes the PageSpeed Insights score matter?
The PageSpeed Insights score matters as a diagnostic, not as a target. Google's ranking systems use Core Web Vitals from real Chrome users, which PageSpeed Insights shows in the field data panel at the top of the report, while the 0 to 100 performance score below it comes from a single simulated test.
This is the most common confusion we hear from founders. Someone runs their homepage through the tool, sees an orange 60 on mobile and assumes the site is failing. Often the field data above that number says the page passes LCP, INP and CLS for real visitors. For a primer on the three metrics, read Core Web Vitals explained for marketers and founders.
What is the difference between lab data and field data?
Lab data is one test run in controlled conditions; field data is what actually happened to real visitors over the past 28 days. PageSpeed Insights shows both on the same page, which is why they're so easy to mix up.
Field data (real users) | Lab data (Lighthouse score) | |
|---|---|---|
Source | Chrome UX Report from real Chrome visitors | One simulated page load by Lighthouse |
Time span | Rolling 28 days | A single moment |
Devices and networks | Your actual visitors' devices and connections | One emulated device with throttled CPU and network |
Metrics | LCP, INP, CLS, plus FCP and TTFB | Performance score built from FCP, LCP, TBT, CLS and Speed Index |
Responsiveness | INP from real interactions | Total Blocking Time as a proxy; no real clicks |
Used by Google for page experience | Yes | No |
Best used for | Judging whether you pass | Finding why a page is slow |
The lab test is deliberately harsh. Lighthouse's mobile test emulates a mid-range phone on a slow connection, which is useful for finding problems but often slower than what most of your visitors experience. Run it twice and the score can change by several points, because of network and server variance.
How can a 60 PageSpeed score still pass Core Web Vitals?
A 60 score can still pass Core Web Vitals because the score measures a simulated worst-ish case, while the vitals assessment measures real visits at the 75th percentile. If your visitors mostly use decent phones and connections, they can get good LCP, INP and CLS even when the throttled lab test looks mediocre.
- Different conditions. The lab simulates slower hardware and network than many real audiences use.
- Different metrics. The score includes Speed Index and Total Blocking Time, which aren't Core Web Vitals.
- No real interactions. The lab can't measure INP, so it estimates main-thread blocking during load instead.
- Caching and repeat visits. Real visitors often have cached assets; the lab always loads cold.
The reverse happens too. A page can score 90 in the lab and fail INP in the field, because a chat widget or tag manager makes real clicks slow after the page has loaded. Lab tests don't click anything, so they miss it. That's why how to fix INP starts with field data, not with the score.
Which number should you act on?
Act on field data first. If the Core Web Vitals assessment at the top says passed for the page or for the whole origin, you don't have a Core Web Vitals problem, whatever the score says.
Read the top panel first
Check the field data assessment for the URL and the origin, on mobile and desktop. Note any metric marked needs improvement or poor.
If field data passes, stop worrying about the score
Use the diagnostics below the score only to find cheap wins, such as oversized images or unused scripts.
If field data fails, find the matching lab clue
Failing LCP points to the LCP element and server response. Failing INP points to long tasks and Total Blocking Time. Failing CLS points to layout shift diagnostics.
If there's no field data, use the lab carefully
Low-traffic pages often have none. Use the lab score as a rough guide and consider real-user monitoring for important pages.
Confirm in Search Console
The Core Web Vitals report groups similar pages, which shows whether a problem is one page or a whole template.
Hypothetical example: three pages, three different stories
For example, suppose a startup checks three pages. The homepage scores 58 on mobile, but field data passes all three vitals: no action needed beyond cheap wins. The pricing page scores 91, but field data shows INP in needs improvement on mobile: the lab missed a slow plan toggle and a chat widget, so that's the page to fix. A new landing page scores 72 and has no field data yet: the team uses the lab diagnostics to fix its hero image and checks again once traffic builds.
Does the PageSpeed score affect rankings or conversions?
The PageSpeed score itself doesn't affect rankings. Google has said its page experience signals use Core Web Vitals field data, and those are a modest signal behind relevance and content quality. The score is a tool output, not something Google ranks on.
Conversions follow real experience, not the score. A visitor who waits for the hero or taps a button that doesn't respond is the one who leaves, and that's what field data captures. We look at what the evidence does and doesn't show in page speed and conversion rate. Google documents how Core Web Vitals relate to Search on Search Central.
How should you use Lighthouse diagnostics well?
Use Lighthouse diagnostics to find the cause behind a failing field metric, then fix that cause and confirm in field data. The opportunities list is a menu, not a to-do list.
- Largest Contentful Paint element: tells you which element to fix; see how to improve LCP.
- Reduce unused JavaScript and third-party usage: often points at tags; our third-party scripts audit explains how to decide what goes.
- Avoid large layout shifts: names the elements moving, which maps directly to CLS fixes.
- Properly size images: a quick win on most sites.
- Platform specifics: builders limit what you can change; see Core Web Vitals for Framer, Webflow and Next.js.
