CWV Measurement: Field vs Lab Data (Which Wins in 2026?)

Field data comes from real users visiting your site. Chrome collects this through the Chrome User Experience Report (CrUX). It aggregates data from opted-in Chrome users across millions of origins. Field data includes metrics like LCP, INP, and CLS as real people experience them.

CrUX data is reported monthly. It shows the distribution of metric values across your origin. The 75th percentile is the pass or fail threshold. Field data reflects actual device capabilities, network conditions, and user behavior. It is the ground truth for ranking eligibility.

As of 2026, CrUX covers over 40 million origins. HTTP Archive uses CrUX data to generate its annual Web Almanac reports. Field data is available through PageSpeed Insights, the CrUX API, and CrUX dashboards in Google Looker Studio.

Lab data runs tests in a synthetic environment. Tools like Lighthouse, PageSpeed Insights lab mode, and WebPageTest simulate page loads on a controlled device and network. Lab data is reproducible and consistent. It helps developers isolate specific performance issues.

Lab data shows detailed diagnostic information. It tells you which element is the LCP candidate, which JavaScript files block rendering, and where layout shifts occur. Field data cannot provide this level of detail. Both types of data serve different purposes.

The standard lab environment for CWV uses a Moto G4 emulation with slow 3G throttling. This is a middle-ground setup that catches most common issues without being overly punitive.

Lab data measures a single, controlled test. Field data measures thousands of real user sessions. Lab data tells you what could happen. Field data tells you what actually happens. Lab data is essential for debugging. Field data is essential for ranking validation.

Field data can vary day to day based on traffic composition. Lab data is stable across runs. Field data captures long-tail issues like slow interactions on low-end devices. Lab data misses these entirely.

A page can pass lab tests but fail field tests. This gap is common. It happens when lab tests use a faster network than real users, or when lab tests miss interactions that field users perform. The opposite scenario, passing field but failing lab, is rare but can occur if lab tests simulate more aggressive throttling than your typical users experience.

Field data always wins for ranking. Google uses CrUX field data to determine whether a page passes or fails each CWV threshold. Lab data is not used for ranking. Lab data is used for debugging and optimization guidance.

This distinction matters for prioritization. If your CrUX data shows all three metrics in the green, you are done for ranking purposes. Keep monitoring, but do not chase improvements in lab scores that already pass in the field. If field data shows failures, use lab data to diagnose the root cause.

Data from the 2025 HTTP Archive shows that 40% of origins have better lab scores than field scores for LCP. This discrepancy highlights the importance of field data as the ranking signal.

Set up CrUX monitoring for your origin using the CrUX API or a dashboard tool. Run weekly Lighthouse CI tests to catch regressions early. Compare lab results against field results monthly. Investigate any metric where lab and field diverge by more than 20%.

Note the gap between your lab and field scores. Lab data alone can give a false sense of confidence. Field data reveals what real devices and network conditions do to your performance. Audit quarterly.

• web.dev/lab-and-field (https://web.dev/lab-and-field-data-differences)
• CrUX Documentation (https://developer.chrome.com/docs/crux)
• HTTP Archive Field vs Lab Analysis (https://httparchive.org/reports/chrome-user-experience#fieldLab)
• PageSpeed Insights Docs (https://developer.chrome.com/docs/psi)