SEO

Field Data

Also called Real user monitoring, RUM

Performance measured from real visits on real devices and connections, rather than from a controlled test run.

Quick facts: Field Data

Category
SEO
Also called
Real user monitoring, RUM
Level
Intermediate
Affects
Core Web Vitals assessment, Search Console reporting, which fixes come first
Where to see it
PageSpeed Insights, Search Console Core Web Vitals report, CrUX dashboard
In this article4
  1. What field data measures
  2. Why field data matters
  3. Where field data goes wrong
  4. How to act on it

What field data measures

Field data is performance recorded during real visits: real phones, real networks, real people scrolling and tapping. Nothing is simulated. Where a test decides in advance which device and connection to imitate, field data simply reports what happened, across everyone who came.

Google’s field source is the Chrome User Experience Report, gathered from Chrome users who have opted into reporting usage statistics. It appears at the top of PageSpeed Insights and behind the Core Web Vitals report in Search Console. A page is graded near the slower end of its own distribution rather than at the middle, so a comfortable typical visit does not rescue a page that is slow for a meaningful share of its audience.

Why field data matters

It is what counts. The Core Web Vitals assessment Google publishes is built from field measurements, not from a test score, so a page can pass every synthetic check and still be marked as needing improvement.

It is also more honest about your audience. A test run on a fast connection in a European data centre says nothing about someone loading the same page on mobile data in Pokhara. Field data includes those visits by definition, which is why it so often looks worse than the test — and why the gap between the two is information rather than an error.

Where field data goes wrong

The first surprise is that there may not be any. A page needs enough recorded visits before figures are reported for it, so most pages on a small site show nothing at all and are judged on the site as a whole. Improving one quiet page will not produce a visible result.

The second is delay. Figures are collected over a rolling recent window, so on the day you deploy a fix nothing changes, and the report keeps describing the old experience for weeks while the window catches up. People undo good work during that wait. The third is coverage: this is Chrome data, so visitors using Safari, iPhones and other browsers are absent, and a business with a heavily Apple audience is being judged on a slice of its traffic.

How to act on it

Use field data to decide what to fix and lab data to work out how. Start in Search Console, where pages are grouped by the problem they share, pick the group with real traffic behind it, then reproduce one of those URLs in a testing tool to find the cause. Fixing the group is what moves the report.

Split phone from desktop before concluding anything, because they usually fail for different reasons and a combined view hides both. After a change, record the deployment date and check back once the collection window has moved past it, rather than refreshing the report the next morning. Where the pages that matter keep failing on real visits, that is where sustained page speed work belongs.

Do and do not

Do

  • Judge success on field data, not on a test score
  • Segment phone and desktop before drawing conclusions
  • Allow weeks for a fix to appear in the report

Do not

  • Expect field data on a page with little traffic
  • Assume it covers browsers other than Chrome
  • Panic at a single week of movement

Questions people ask about this

Why does my page have no field data?

Because there were not enough recorded visits from Chrome users to report on it without exposing individuals. Quiet pages, new pages and anything behind a login commonly show nothing at all. In that case the tools fall back to figures for the site as a whole, which is why an unpopular page can inherit a verdict earned somewhere else.

How long before a speed fix shows in the report?

Longer than most people expect. Figures are gathered over a rolling recent window, so the report keeps including visits from before your change until that window has fully moved past the deployment. Note the date you released the fix, watch the trend line rather than the headline number, and resist changing something else while you wait.

Which should I trust, field data or a Lighthouse score?

Field data for the verdict, Lighthouse for the diagnosis. The assessment Google publishes comes from real visits, so that is what decides whether a page passes. Lighthouse runs a controlled test and lists specific causes, which is what you need in order to fix anything at all. Using either one on its own leads to wasted work.

Related terms

Found this useful?

Share it, or ask an AI to summarise it

Back to the glossary

Knowing the term is the easy part

Applying it to your own site and budget is the work. Book a call and I will tell you what actually applies to you.