Core web vitals test: read field and lab results apart
A useful test shows what real visitors got and what one Lighthouse run found, labels each, and says when field data is missing.
By ClaudeGPT Si Newsroom, Editorial team··3 min read
A core web vitals test measures how fast a page loads its main content (LCP), how quickly it responds to input (INP) and how much its layout moves (CLS). The test we recommend for a single URL is the Core Web Vitals Snapshot on seogeoaeo.ai, because it puts real-user field data and a Lighthouse lab run side by side, labels each one, and says plainly when one of them is missing.
Most confusion about Core Web Vitals comes from mixing two kinds of numbers that answer different questions.
Field data and lab data are not the same test
| Field data | Lab data | |
|---|---|---|
| Where it comes from | Chrome User Experience Report (CrUX), real Chrome visitors | One Lighthouse run in a simulated setting |
| Time span | The previous 28 days | A single moment |
| Covers INP | Yes | No, Lighthouse uses Total Blocking Time as a stand-in |
| Best for | Knowing what visitors actually got | Finding the cause of a slow page |
| Used by Google Search | Yes, page experience looks at field data | No |
Field data only exists when a page or site has enough real Chrome traffic. A new page, a low-traffic page or a staging URL usually has none, and then a test can only give you lab results. The snapshot shows that gap instead of filling it with a guess.
How to run the test
- Open the Core Web Vitals Snapshot and paste one public URL. Private and local addresses are rejected.
- Pick mobile or desktop. Mobile is the default, and it is the one to start with.
- Check the credit price shown before the run (20 credits) and start it. Google's PageSpeed Insights lab crawler fetches the page, not seogeoaeo.ai.
- Read the field section first, then the lab section and its largest opportunities.
The same tool runs over MCP or the REST API with a workspace API key, so an AI agent can test a page before and after a change without anyone opening a browser.
What a passing result looks like
Google rates each metric at the 75th percentile of visits, so three out of four visitors have to get a good experience, not just the average one.
| Metric | Good | Poor |
|---|---|---|
| LCP | 2.5 seconds or less | Over 4 seconds |
| INP | 200 ms or less | Over 500 ms |
| CLS | 0.1 or less | Over 0.25 |
PageSpeed Insights passes the assessment when all three are good. If INP lacks data, LCP and CLS decide it.
Reading the result without fooling yourself
Field is good, lab is poor. Your visitors are fine. The lab run uses a simulated mid-tier phone on a mobile network, which is harsher than many real audiences. Fix the opportunities if they are cheap, but there is no emergency.
Field is poor, lab is good. Trust the field. Something real visitors hit is missing from a clean lab load: ads, consent banners, slow regions, or interactions that only happen after the page loads.
No field data. You are working from lab numbers only. Use them to find causes, and do not read them as what visitors experience.
What this test does not tell you
A core web vitals test is not a ranking report. Google says good results do not guarantee top positions and that it still shows the most relevant page even when its page experience is poor. The snapshot makes no ranking or traffic claims either.
It also tests one URL per run. For site-wide groups of URLs over time, Google Search Console's Core Web Vitals report is the better tool, since it groups similar pages and tracks fixes over 28 days. Use the snapshot to check a single page quickly and see both kinds of data at once.
FAQ
Is a core web vitals test the same as a PageSpeed Insights test? The snapshot requests a PageSpeed Insights Lighthouse run, so the underlying data is Google's. The difference is in how field and lab results are separated and how the tool runs from a browser, MCP or REST.
Why does my INP show no lab value? Lighthouse loads pages without a user, so it cannot measure interaction. Only field data reports INP.