Lab data vs field data: which speed numbers to believe

Key takeaways

Lab data is a speed test run once on a simulated device. Field data is what real visitors experienced over the past 28 days. When they disagree, believe field data about how your site is doing and use lab data to work out why. Google's ranking systems use field data. Most reports sent to business owners quote the lab score, which is why the two are so often confused.

• Lab: one simulated visit, repeatable, available for any page.
• Field: thousands of real visits, on real phones and connections.
• Lab: explains causes and lets you test a fix immediately.
• Field: shows the outcome, and takes 28 days to reflect a change.

What is the difference between lab and field data?

Lab data is performance measured in a controlled test, using a fixed device and network setting. Field data is performance recorded from real users as they visit, on whatever device and connection they have.

In PageSpeed Insights, the top section labelled as what real users are experiencing is field data from the Chrome User Experience Report. The lower section, with the performance score and diagnostics, is a lab test run by Lighthouse.

• How website speed affects customer experience and conversions
• How to test website speed the way customers feel it

Why do lab and field results disagree?

They disagree because they measure different visits. A lab test loads the page once, from the top, with nothing stored, on a simulated mid-range phone, and does not scroll or tap. Real visitors have a range of phones and connections, often have parts of your site stored already, and interact with the page.

• Real devices may be faster or slower than the simulated one.
• Returning visitors load less, so field loading can look better.
• Field data includes taps and scrolling, which reveal slow responses and late shifts.
• Cookie banners and personalised content appear for some visitors and not others.
• The lab test runs from a data centre, not from your customers' locations.

• Why do website speed tests give different results?

Which one does Google use?

Google uses field data. The Core Web Vitals that its ranking systems consider are measured from real Chrome users and judged at the 75th percentile, meaning three quarters of visits must meet the threshold. The lab performance score is not a ranking input.

This matters when someone promises a score of 95. A high lab score on a site whose real visitors have slow phones and a heavy consent banner can still fail in the field.

• Core Web Vitals for business websites
• Does website speed affect SEO and Google rankings?

How should a business use each?

Use them for different jobs. Field data is the scoreboard. Lab data is the workshop.

• 1. Check field data to see whether real visitors have a problem.
• 2. Use lab diagnostics to find what is causing it.
• 3. Make a change and re-run the lab test to confirm it helped.
• 4. Wait 28 days and check field data to confirm visitors felt it.
• 5. Report field results to the business, and lab findings to the developer.

• A website speed report directors will actually read
• What is a good PageSpeed Insights score for a business?

What if your site has no field data?

Small sites often have none, because the Chrome User Experience Report only publishes results where enough visits have been recorded. PageSpeed Insights may show data for the whole site when a single page has too little, or nothing at all.

In that case, lab results and real-phone testing are your evidence. You can also collect your own field data by adding a small measurement script, though for most small businesses a quarterly test by hand is enough.

• Site speed in GA4: how to connect speed to enquiries

What does this look like in practice?

A pattern we see in agency reports: a green lab score for the homepage is presented each month as proof the site is fast. Search Console, which nobody opened, shows most mobile pages as needing improvement in the field, because real visitors meet a late-loading banner and slow taps the lab test never triggers.

Reporting field results first, and using the lab only to explain them, changes which work gets done. Global Bridge Labs (GBL) reports both and labels them clearly.

• See client work
• Website Design & Development for UK businesses

Lab and field data checklist

• Find the field data section for each key page.
• Note LCP, INP and CLS and whether each passes.
• Use the lab diagnostics to list causes.
• Re-run the lab test after each fix.
• Re-check field data after 28 days.
• Ask any supplier which kind of data their report shows.

• The website speed scorecard: score your site in an hour

Next step

If you have only ever been shown one speed number, you have probably been shown the wrong one. We will pull both for your site and explain what they mean together.