Lab data vs field data: why your speed scores disagree
Definition
Lab data is a measurement from a single controlled test run on a simulated device and connection. Field data is collected from real visitors over time. They often disagree, and field data is the better guide to what your customers experience.
Why they differ
A lab test uses one device profile, one location and one moment. Real visitors arrive on many phones, on good and bad connections, with logged-in states and warm caches. Averaging all of that gives a different picture than one simulated run.
What each is good for
Lab data is repeatable, so it is ideal for testing a change: run it before and after and compare.
Field data shows what really happened, including problems a lab run never triggers. It also cannot be produced on demand, so it is slow to react to a fix.
When there is no field data
Low-traffic pages have no field data because Google needs enough visits to report a figure. That is not a fault. Rely on lab data for those pages and treat it as an estimate.
How the audit uses both
The audit reads both from PageSpeed Insights and prefers field data when it exists. The report says which one it used.
Frequently asked questions
Which should I fix against?
Use field data to decide whether there is a problem and lab data to test fixes.
Why did my lab score change with no code change?
Lab runs vary by a few points because of server load, network noise and third-party content.