See field CrUX and lab Lighthouse as two different measurements.

What this check actually measures

Page speed is two measurements, not one. Field CrUX is what real Chrome users experienced on this URL (or, if the URL is too thin, on the origin). Lab Lighthouse is one simulated mobile run: a performance score and the opportunities that run named. This check calls PageSpeed Insights for mobile and keeps those two stories apart. It does not blend them, and it does not turn missing field data into a zero.

The failure modes, and what each one means

Not enough real-user data

CrUX needs enough Chrome traffic. When it is absent, the field block is null — not 0 ms, not a green score. The lab run may still complete. Absence is a measurement gap, not a pass.

A poor or needs-improvement band

LCP, INP, and CLS are reported with the CrUX bands Google publishes. A poor band is what users are hitting, not a lab guess. Fix the named metric; this check does not rank the page.

A quota or API miss

PageSpeed Insights is free and quota-bound. A rate-limit or empty response is a failed check, not a speed score. Retry later; nothing is invented to fill the gap.

Why it matters for search — and for AI answers

Search uses field Core Web Vitals where they exist. A lab-only number is a different object: useful for debugging, not a substitute for CrUX. Answer engines that fetch your page still wait on the same render path users wait on. This check reports the live PSI payload. It does not claim a ranking or citation effect.

What the full product adds

This tool reads one URL, once. The TrustGrowth audit samples PageSpeed on a schedule, files slow LCP and high CLS against the pages they affect, and keeps the evidence next to the score.