Field data and lab data are the two ways of measuring web performance. Field data, also called real user monitoring, is collected from actual visits on real devices and networks. Lab data, or synthetic testing, comes from loading a page in a controlled environment such as Lighthouse. Google's Core Web Vitals assessment uses field data, while lab data is the main tool for diagnosing and testing fixes.
How they differ
Lab tests use one device profile, one network speed and one location, with no user interaction and often an empty cache. Real users bring a wide range of phones, connections, cached states, logged-in conditions and behaviors. The two rarely match, and that is expected.
- Field: real conditions, reported as a distribution, often at the 75th percentile
- Lab: repeatable, detailed traces, good for comparing before and after
- Field captures interactions, so it can measure INP; lab mostly cannot
- Lab works before launch; field needs traffic
Using both
Start with field data to find which page types have problems for real users. Then use lab tools to reproduce and diagnose the cause, and to check that a fix helps before release. Confirm the result in field data after release. If the lab says a page is fast but the field says it is slow, look at conditions the lab does not simulate: slower devices, interactions, third-party scripts that vary, or geographic distance.
Sources of each
Field data comes from the Chrome UX Report, Search Console and your own real-user monitoring. Lab data comes from Lighthouse, PageSpeed Insights' lower section, WebPageTest and the Chrome DevTools performance panel.
Example
A fast lab score and slow real users
Copperloom's homepage scores well in Lighthouse, yet CrUX shows a poor LCP. Its own monitoring splits the field data by country and finds the problem concentrated far from the single origin region. Putting HTML behind a CDN cache brings field LCP in line with what the lab showed.
Common mistakes
Reporting lab scores as user experience
A lab score describes one simulated load. Use field data when you talk about what users experience.
Ignoring lab data because it differs
Lab tools are where you find causes. The mismatch is a clue, not an error.
Questions
Google's page experience documentation refers to Core Web Vitals measured from real-user field data.
