Why has PageSpeed Insights now incorporated real user data in addition to simulated scores?

What’s the difference between simulated data and real user data?

Simulated performance scores (also known as Lab data) are generated in a controlled environment: stable connection, standardized device, empty cache. It's reproducible, but it doesn't necessarily reflect the experience of your real visitors.

Real user data (Field Data or CrUX) comes from millions of Chrome users who have agreed to share their browsing stats. It captures real-world conditions: degraded mobile networks, low-end devices, varied geolocation. It’s raw, sometimes noisy, but it represents the reality of your performance.

Where exactly does this real data come from?

Google collects these metrics through the Chrome User Experience Report (CrUX). Only users who have explicitly enabled the sharing of usage statistics in Chrome contribute to this dataset.

CrUX aggregates data over a rolling 28-day period, by origin (full domain) and sometimes by individual URL if traffic is sufficient. The metrics include LCP, FID, CLS, TTFB, and other Core Web Vitals measured under real conditions.

Why is Google adding this layer now?

Because Core Web Vitals have become an official ranking factor. Measuring only in the lab is no longer enough: Google wants you to optimize for the actual experience, the one that matters for ranking.

PageSpeed Insights thus becomes a unified dashboard: you see your theoretical performance (what you can achieve in optimal conditions) alongside your real-world performance (what your users actually experience). The gap between the two often reveals hidden issues: network problems, geographic variability, unstable third-party resources.

• Lab Data: reproducible, ideal for diagnosing and testing optimizations in development
• Field Data (CrUX): reflects the real experience that impacts Google ranking
• Lab/Field Gap: often indicates specific issues affecting certain user segments (3G mobile, remote geographies)
• 28-Day History: CrUX data is not instantaneous; it takes time to see the impact of an optimization
• Traffic Threshold: low-traffic URLs do not have individual CrUX data, only at the origin level

Is this Lab + Field integration really new?

No, and it's important to emphasize that. CrUX has been around since 2017, and PageSpeed Insights has been displaying it for several years. What’s changing here is likely the clearer emphasis on these two data sources in the interface.

The real novelty is not technical but communicational: Google is now insisting that webmasters stop focusing solely on the Lab score. Too many sites optimize for a 100/100 in the lab while delivering a poor experience in real conditions. [To be verified] whether this update involves a specific interface redesign or simply a reminder of best practices.

Are CrUX data reliable for decision-making?

Yes, but with nuances. CrUX relies on a biased sample: only Chrome desktop and Android (not iOS), only users who have enabled data sharing (often more tech-savvy). Safari, Firefox users, or those who block telemetry are not counted.

For high-traffic sites with diverse audiences, this bias is diluted. For specific niches (tech-savvy users on Firefox, iOS-heavy audiences), CrUX may under-represent your real issues. In these cases, supplement with your own RUM tools (Real User Monitoring) to capture 100% of your audience.

Should you really aim for 100/100 in both Lab and Field?

No. Aiming for 100/100 in Lab is often counterproductive: it leads to costly micro-optimizations for marginal gains. What matters for ranking is to reach the