Search results and measurement
PageSpeed Insights: meaning, examples and practical use
PageSpeed Insights is Google’s performance analysis interface that combines a Lighthouse laboratory test for a URL with CrUX real-user data when available. These sections use different sources and periods and should therefore be interpreted separately.
Also known as: PageSpeed Insights
At a glance
- Real-user observations and the lab report are different measurements.
- Check whether the report shows URL-level or origin-level data.
- A high score does not guarantee rankings or every visitor’s experience.
On this page

What does the report contain?
The PageSpeed Insights documentation identifies CrUX as its field-data source and Lighthouse as its diagnostic test. The former aggregates earlier visits; the latter reports a controlled run. Missing field observations do not make the lab report unusable.
Why can the sections disagree?
Devices, networks, time windows and observed interactions differ. A new fix can already appear in a lab run while older visits still influence the field summary. That difference alone does not establish a measurement error.
Example: after image optimization
Imagine a page receiving a smaller main image. A new laboratory run measures better LCP while the real-user summary still includes earlier days. The team records the change date and later checks relevant field data rather than immediately claiming improvement for every visitor.
How should an action be selected?
Use diagnostics to identify the affected resource. Do not apply every general suggestion blindly. A performance budget can help prevent regressions, but assess the report alongside the visitor's actual task. A faster score with a broken action button is not a successful overall change.