Page Speed Test: Lab vs Field Data

A useful speed test separates one simulated run from the trailing 28 days of real-user evidence. Each layer answers a different question.

Run a test: Foglift's free Page Speed Checker returns Lighthouse lab metrics and URL-level CrUX percentiles when Google has sufficient field data.

Run the test in a fixed sequence

Test the exact canonical URL. Redirects and alternate templates can produce different request paths.
Record mobile and desktop separately. Lighthouse uses different scoring curves and conditions.
Save the environment. Keep the test date, location, device profile, and cache state.
Repeat the lab run. Chrome documents variation from routing, hardware, extensions, ads, and A/B tests.
Read field coverage first. Confirm whether CrUX reports the URL, falls back to the origin, or is unavailable.

Lab and field data have different jobs

Evidence	Measures	Best use	Boundary
Lighthouse lab	One simulated page load	Reproducing regressions and inspecting diagnostics	Can vary and miss real-user bottlenecks
CrUX field	Eligible Chrome experiences over the previous 28 days	Real-user distributions and Core Web Vitals status	Needs sufficient traffic and may fall back to origin data
First-party RUM	Events from the audience you instrument	Segmenting by template, release, device, and geography	Requires deliberate sampling and privacy controls

Interpret field metrics at the 75th percentile

PageSpeed Insights reports the 75th percentile for field metrics. A page or origin passes when the available LCP, INP, and CLS values all land in the Good range.

Metric	Good	Needs improvement	Poor
LCP	≤ 2.5s	> 2.5s and ≤ 4.0s	> 4.0s
INP	≤ 200ms	> 200ms and ≤ 500ms	> 500ms
CLS	≤ 0.1	> 0.1 and ≤ 0.25	> 0.25

Read the Lighthouse score as a diagnostic summary

Lighthouse labels 90 to 100 Good. Chrome documents that this weighted score can fluctuate because of routing, hardware, extensions, ads, and experiments. Record the raw metrics and trace evidence alongside it. Compare repeated runs using the same configuration.

A green lab score with poor field data points to conditions or user segments the simulation missed.
A poor lab score with good field data can still reveal a reproducible regression.
A changed score without changed raw metrics can reflect a Lighthouse version or weighting change.

Keep speed, ranking, and AI visibility separate

Google says Core Web Vitals are used by ranking systems. Its guidance also says there is no single page-experience signal and strong tool results do not guarantee top rankings. A speed test supports technical diagnosis; it cannot predict a rank change.

Speed results do not measure whether an AI engine mentions or cites a brand. Use an AI Visibility Check for prompt-level outcomes and the broader Technical Audit for page-level SEO, performance, security, accessibility, and AI Readiness evidence.

Turn the result into a release decision

Name the failing metric and affected URL or template.
Attach the trace, waterfall, or field segment that supports the diagnosis.
Set a metric budget for the same test configuration.
Verify the fix in repeated lab runs.
Watch RUM or CrUX to confirm that real-user percentiles move.

For implementation techniques, use the existing site speed optimization guide.

Frequently asked questions

What is a good PageSpeed Insights score?

Lighthouse labels 90 to 100 Good, 50 to 89 Needs Improvement, and 0 to 49 Poor. That score describes one controlled lab run. Use it to debug regressions, then check CrUX field data before drawing conclusions about real users.

What is the difference between lab and field data?

Lab data is a simulated Lighthouse load under controlled conditions. PageSpeed Insights field data comes from eligible Chrome users in CrUX and represents the previous 28-day collection period. Lab data is diagnostic; field data describes observed user experience.

Why does PageSpeed Insights show no field data?

CrUX requires enough eligible real-user samples for the tested URL. PageSpeed Insights can fall back to origin-level data when URL coverage is insufficient. If the origin also lacks enough data, the report shows lab results without a field assessment.

Does a high Lighthouse score guarantee better rankings?

No. Google says its ranking systems use Core Web Vitals, but there is no single page-experience signal and good tool scores do not guarantee a top ranking. Relevance and other ranking systems still matter.

Which Core Web Vitals thresholds should I use?

At the 75th percentile, Good means LCP at or below 2.5 seconds, INP at or below 200 milliseconds, and CLS at or below 0.1. A page or origin passes when the available Core Web Vitals meet their Good thresholds.

Sources & Further Reading

Fundamentals: Learn about GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) (the two frameworks for optimizing your content for AI search engines).