A common frustration goes like this. You run your site through PageSpeed Insights, see a bright green score, and then open Google Search Console to find the same pages marked as poor. Or the reverse. Which number is real?
Both are real, but they measure different things. Understanding the difference will save you hours of chasing the wrong target.
What Lab Data Is
Lab data comes from a controlled test. Lighthouse loads your page once, usually on an emulated mid-range phone with a throttled network and CPU, and records what happens. That is what produces the big performance score from 0 to 100.
Lab data has real strengths:
- It is repeatable, so you can test a change right away.
- It gives detailed diagnostics, such as which script blocks rendering.
- It works for pages with no traffic or pages that are not live yet.
Its weakness is that it is a single simulated visit. It does not know your visitors’ devices, networks or behavior.
What Field Data Is
Field data, also called real user data, comes from the Chrome User Experience Report, or CrUX. It collects performance measurements from real Chrome users who visit your pages, aggregated over a rolling 28-day period. In PageSpeed Insights it appears at the top under the heading about what your real users are experiencing, and in Search Console it feeds the Core Web Vitals report.
This is the data Google uses to evaluate Core Web Vitals. It reflects real devices, real connections, real cookie banners and real third-party scripts. Its weakness is that it is slow to change, and it needs enough traffic to exist.
Why the Two Disagree
There are several common reasons:
- Different devices. Your visitors may use slower phones or weaker networks than the lab simulates, or faster ones.
- Different pages. Field data is grouped across many URLs, while a lab test covers one page.
- Different interactions. Lighthouse does not measure INP, because it does not click or type. Real users do.
- Different moments. Lab tests do not scroll the page the way a person does, so layout shifts that happen later may be missed.
- Different timing. Field data reflects the last 28 days, so a fix you made yesterday will not show up yet.
- Logged-out, cache-cold test vs. real visitors, many of whom have cached files from earlier visits.
Which One Should You Trust?
Trust field data to decide whether you have a problem. Use lab data to work out how to fix it.
Google’s ranking systems use field data, so a perfect Lighthouse score does not guarantee that you pass Core Web Vitals. I have seen sites with a score of 98 that fail INP in the field because of a chat widget that only misbehaves when real people tap around. I have also seen sites with a score in the 60s that pass comfortably because their audience uses fast devices and the pages are simple.
A Practical Workflow
- Check Search Console. Open the Core Web Vitals report and note which groups of URLs are poor on mobile and desktop.
- Identify the failing metric. Is it LCP, INP or CLS?
- Reproduce it in the lab. Use PageSpeed Insights, Lighthouse and Chrome DevTools to find the cause. For INP, interact with the page yourself since Lighthouse will not.
- Fix one thing at a time. Note the lab result before and after.
- Wait for the field data. After 28 days, confirm that the change made it into the real-user numbers.
What If There Is No Field Data?
If a page does not have enough traffic, PageSpeed Insights may show only origin-level data for the whole site, or nothing at all. In that case, use lab data as your guide and focus on fundamentals: fast server response, optimized images, minimal blocking scripts and reserved space for content.
Stop Chasing 100
A perfect score is not the goal. Lighthouse scores can move several points between runs and between versions of the tool. Spending days to climb from 92 to 100 usually does nothing for your rankings or conversions. Passing all three Core Web Vitals for your real visitors, and keeping the site fast as you add content, is what pays off.
That is the principle behind how I work: field data first, lab data to diagnose, and safe changes that keep your design intact. If you want to see what your real-user numbers look like, request a free audit. If the metrics themselves are still unclear, start with the Core Web Vitals guide for business owners.
