Field data.
Field data is the sum of performance measurements experienced by real users on real devices and real networks.
It is the opposite of lab measurement: an auditing tool opens the page in its own environment assuming a fixed device and a fixed network; field data shows the distribution of real use.
The two diverge routinely, and usually in the same direction: lab good, field poor. The reason is that a substantial share of real users browse on low-powered devices and variable connections.
This is the data used in ranking. A verdict of “we are good” can therefore only be given from field data; a lab score is a diagnostic instrument, not a target.
- Delayed — a twenty-eight-day window; the effect of today’s fix shows weeks later.
- Requires a threshold — no data is collected for sites without enough traffic.
- Aggregated — it may be reported for groups of similar pages rather than page by page.
The way past all three limits is to set up your own measurement: the browser reports all three metrics to the page, and those values can be sent to your own system. It requires no cookies and collects no personal data.
The place where the two measurements DIVERGE is itself a diagnostic: if lab is good and field is poor, the cause is usually the device and network distribution, and the treatment is reducing JavaScript. If both are poor, the diagnosis is made in the lab.
The practical rhythm is this: field data is read monthly and looked at for general health; lab measurement runs on every deployment and catches regressions. The two work at different rhythms because they answer different questions.
On this site field data is pulled into a cache file and shown on screen — but if there is no data the field is not rendered at all. The suite measures both directions: if a value exists the number on the strip must match the cache exactly, and if not the word “measured” must never appear.
- “What is slow?” — lab. Repeatable, detailed.
- “Did the change work?” — lab. Comparison under identical conditions.
- “What are our users experiencing?” — field. The only honest answer.
- “How does Google see us?” — field. The data used in ranking.
- “How is our INP?” — field. It cannot be measured reliably in a lab.
The fifth line is critical: the interaction responsiveness metric looks by definition at real user interactions, and an auditing tool does not interact with the page.
And a note on scope: field data is collected only for pages that receive real visits. A page just published or never visited does not appear in that data at all, however technically flawless it may be.