Speed & performance · Explainer
Lab Data vs Field Data: Why Do Your Scores Disagree?
Short answer
Lab data comes from a single simulated page load in Lighthouse, while field data comes from real Chrome users visiting your page over the previous 28 days, reported by the Chrome User Experience Report. The two disagree because they measure different things: one controlled test against thousands of real visits on many devices and networks. Use field data to judge how fast your site feels, and lab data to find what to fix.
Key takeaways
- Lab data is one controlled Lighthouse run; field data is real Chrome visits over 28 days.
- Google judges Core Web Vitals at the 75th percentile, separately for mobile and desktop.
- A page passes only when LCP, INP and CLS are all good: 2.5 seconds, 200 milliseconds and 0.1.
- Lighthouse cannot measure INP, so it uses Total Blocking Time as a stand-in.
- Field data takes up to 28 days to show a fix, because old visits stay in the window.
- A Lighthouse score of 90 or more is good, but it is a lab score, not the Core Web Vitals verdict.
What is lab data?
Lab data is the result of loading your page once in a controlled setting with a fixed device, network and location.
Chrome tools that report lab data generally run Lighthouse, the open-source audit tool built into Chrome. Lighthouse loads the page, runs a set of audits, and reports scores for Performance, Accessibility, Best Practices and SEO.
The point of a lab test is control. The same page, tested the same way twice, should give close to the same result, so you can tell whether a change helped or hurt.
A lab run also starts with an empty cache, like a first-time visitor. It does not know what your returning visitors already have stored, or how they behave on the page.
What is field data?
Field data is the set of measurements taken from real people visiting your page on their own devices and networks.
It is also called Real User Monitoring, or RUM. web.dev explains that field data reflects the actual devices, network conditions and geographic locations of your users, and you can read the full comparison in Why lab and field data can be different.
In Google tools, field data comes from the Chrome User Experience Report, known as CrUX. Search Console Help says the CrUX report gathers anonymized metrics about performance times from actual users visiting a URL, and its Core Web Vitals report is built on that data.
Field data is not one number. It is a distribution of thousands of visits, from fast to slow, and Google reports it at the 75th percentile.
The Web Almanac 2025 Performance chapter combines HTTP Archive lab measurements with CrUX real-user data. It reports 48% of mobile sites and 56% of desktop sites passed all Core Web Vitals in 2025.
Why do lab and field scores disagree?
They disagree because they measure different things: one controlled load against thousands of real visits on varied devices, networks and locations.
| Feature | Lab data | Field data |
|---|---|---|
| Source | One Lighthouse run | Chrome User Experience Report (CrUX) |
| Time window | A single moment | Previous 28 days |
| Device | One simulated device | Every visitor's device |
| Network | One fixed connection | Every visitor's connection |
| Cache | Empty, like a first visit | Mixed, many returning visitors |
| INP | Not measured, Total Blocking Time stands in | Measured from real taps and clicks |
| Best use | Debugging a problem | Judging the result |
A lab test runs on one device, one network and from one place, while your visitors arrive on old and new phones, fast fiber and weak mobile signals, from many countries. That gap alone can move every metric.
Cache state adds a second gap. The lab starts cold, but returning visitors may load most files from their browser cache, so field data can look better than the lab. The reverse happens when visitors use slower phones than the lab simulates.
What does the 75th percentile mean for your scores?
The 75th percentile is the value that 75% of visits meet or beat, so a page passes only when at least 3 in 4 visits are fast enough.
Picture 100 visits sorted from fastest to slowest. The 75th visit sets the score, so a few very fast visits cannot hide a slow tail. web.dev recommends measuring the 75th percentile of page loads, split by mobile and desktop, in its Web Vitals guide.
The good thresholds, checked November 2026, are the same on mobile and desktop, but each device type is judged on its own.
| Metric | Good | Needs improvement | Poor |
|---|---|---|---|
| LCP | 2.5 seconds or less | Over 2.5 up to 4 seconds | Over 4 seconds |
| INP | 200 ms or less | Over 200 up to 500 ms | Over 500 ms |
| CLS | 0.1 or less | Over 0.1 up to 0.25 | Over 0.25 |
A page passes only when all three metrics are good at the 75th percentile. One weak metric fails the whole assessment, even if the other two are green.
If you want the targets translated into plain WordPress terms, read how fast a WordPress site should load.
Why does the 28-day window delay your results?
Field data covers a rolling 28-day window, so a fix only shows fully once 28 days of faster visits have replaced the older, slower ones.
Chrome's CrUX API documentation states that the data in the report is a 28-day rolling average of aggregated metrics, and that it is updated daily around 04:00 UTC.
That delay catches many site owners out. You can fix a slow hero image on Monday and see a better lab score the same day, yet the field number barely moves for weeks.
Google's Core Web Vitals report help says to click Start Tracking to begin a 28-day monitoring session after a fix. The report groups similar URLs together, so it shows your site as a whole rather than one exact page.
Why can a page pass Lighthouse and still fail Core Web Vitals?
A page can pass the lab and fail the field because the lab never sees real taps, real scrolling or real phones, and it scores a different mix of metrics.
Interaction to Next Paint, or INP, measures how fast the page shows a response after a tap, click or key press. A lab load has no user, so it cannot measure INP, and uses Total Blocking Time as a stand-in instead.
For the full definition and the 200-millisecond good threshold, see what INP is.
Largest Contentful Paint can also differ. The largest element in the lab may not be the one real visitors see, because screen size, personalized content and font loading can change it.
Cumulative Layout Shift differs too. The lab counts shifts only during loading, while field data counts shifts over the whole visit, including late banners and content that loads after scrolling.
If layout shift is your failing metric, our guide to fixing cumulative layout shift in WordPress walks through the common causes.
INP became a Core Web Vital on March 12, 2024, replacing First Input Delay, according to web.dev. Any report or guide that still judges responsiveness by FID is out of date, as of November 2026.
Why can field data look better than lab data?
Field data can beat the lab when most visitors are returning readers with cached files, fast phones or nearby servers.
The lab always starts cold with an empty cache. A returning visitor may load your page mostly from the browser cache, and someone pressing the back button may get the page almost at once from memory.
Geography cuts both ways. A lab test runs from one place, so visitors near your server look fast and visitors far away look slow. Field data blends them all, which is why one lab number never tells the full story.
This is also why a low-traffic page may show no field data at all. Chrome's CrUX methodology page says a page must be publicly discoverable and sufficiently popular to be included, and you cannot submit a page by hand.
What does the 0-100 Lighthouse score actually measure?
The 0-100 Lighthouse Performance score is a weighted average of five lab metrics, not a Core Web Vitals verdict.
Chrome for Developers publishes the weights in its Lighthouse performance scoring guide, set in Lighthouse 10.
| Metric | Weight in the score |
|---|---|
| Total Blocking Time | 30% |
| Largest Contentful Paint | 25% |
| Cumulative Layout Shift | 25% |
| First Contentful Paint | 10% |
| Speed Index | 10% |
Total Blocking Time carries the most weight at 30%, so heavy JavaScript during load can drag the score down even when images are fine. INP is not in the mix at all, because the lab cannot measure it.
The score bands come from the PageSpeed Insights docs: 90 or above is good, 50 to 89 needs improvement, and below 50 is poor. The same page in the About PageSpeed Insights guide says lab data is useful for debugging but may not capture real-world bottlenecks.
First Contentful Paint has its own bands in that guide: good up to 1,800 milliseconds, needs improvement up to 3,000 milliseconds, and poor above that. FCP is a supporting metric, not a Core Web Vital.
How do you check both kinds of data for your WordPress site?
Open PageSpeed Insights, enter the page URL, and read the field section at the top first, then the Lighthouse lab section below it.
- Run the URL in PageSpeed Insights and note the Core Web Vitals verdict and the 75th percentile values for mobile and desktop.
- Open Search Console, go to Experience, then Core Web Vitals, and look at the URL groups marked Poor or Need improvement.
- Run the same URL through Lighthouse again after each fix, and run it a few times, since one lab run can vary.
- Wait for the field data to move, and check the CrUX API if you want the raw 28-day numbers for a page or origin.
PageSpeed Insights shows field data only when the page has enough traffic. If the field section is empty, the report falls back to origin data for the whole site, or shows lab data only.
Once you know which metric fails, follow how to pass Core Web Vitals in WordPress for the fixes in order.
Which number should you trust when they disagree?
Trust field data to judge the result and lab data to find the cause, because only field data shows what real visitors felt.
web.dev puts it plainly: field data should set your priorities, while lab data helps you debug a problem and test a fix before release. Read the lab and field data guide for the full reasoning.
That order matters for rankings too. Google's search systems use field Core Web Vitals, not the lab score, so a green Lighthouse circle does not by itself change how a page is assessed. For the wider picture, see whether page speed affects SEO.
What mistakes should you avoid?
Do not treat the Lighthouse score as the verdict, do not chase one lab run, and do not expect field data to move the day after a fix.
- Judging a page by the 0-100 score alone. The score is a lab blend that leaves INP out, while the verdict uses field LCP, INP and CLS.
- Running Lighthouse once and acting on it. One run can vary, so repeat the test before you undo work that may have helped.
- Checking field data the day after a fix. The 28-day window means old visits stay in the numbers for weeks.
- Fixing the wrong metric first. Start with the metric that fails in the field, not the one with the biggest lab warning.
- Ignoring one device type. Mobile and desktop are judged separately, so check both tabs in PageSpeed Insights.
Frequently asked questions
Can a page score 100 in Lighthouse and still fail Core Web Vitals?
Yes. The Lighthouse score is a lab blend of five metrics that leaves INP out, while the Core Web Vitals verdict uses field data for LCP, INP and CLS at the 75th percentile. A page can look perfect in one simulated load and still be slow for real visitors on older phones, weak networks or long sessions. Check the field section at the top of PageSpeed Insights to see the real verdict.
How long does it take for field data to show a fix?
Up to 28 days. Field data is a rolling 28-day average, so older visits stay in the window until they age out. You should see gradual improvement as new, faster visits replace old ones, and the full effect only lands after about 4 weeks. Lab scores improve at once, which makes them the better tool for checking a fix.
Why does PageSpeed Insights show no field data for my page?
The page has too little traffic to be included in the Chrome User Experience Report. A page must be publicly discoverable and sufficiently popular before Google publishes field data for it. Until then, use the origin data for your whole site if it is shown, and rely on Lighthouse lab results to find and test fixes.
Does lab data affect my Google rankings?
No, not directly. Google's ranking systems use field Core Web Vitals from real users, not the Lighthouse lab score. Lab data still matters, because it shows you what to fix and lets you test changes before release. Treat a good lab score as a useful check, not as proof that visitors have a fast experience.
Should I check mobile or desktop field data?
Both. Google judges mobile and desktop separately, and a page can pass on one and fail on the other. PageSpeed Insights lets you switch between the two at the top of the report, so read each set of 75th percentile values on its own before you decide what to fix.
Sources
- Why lab and field data can be different (and what to do about it) web.dev
- Web Vitals web.dev
- About PageSpeed Insights Google for Developers
- CrUX methodology Chrome for Developers
- CrUX API Chrome for Developers
- Core Web Vitals report Search Console Help
- Lighthouse performance scoring Chrome for Developers
Is your product missing, or something out of date?
Tell us what to add or correct. We read every note and reply when we need a source. A message does not buy a ranking.
Thanks — we review requests within 3 business days.
