Three metrics, three questions
Core Web Vitals measure user experience from three angles. The thresholds below are the ones published on web.dev:
| Metric | What it measures | Good | Needs improvement | Poor |
|---|
| LCP (Largest Contentful Paint) | How quickly the main content appears | ≤ 2.5 s | 2.5–4 s | > 4 s |
| INP (Interaction to Next Paint) | How quickly the page responds to clicks, taps and key presses | ≤ 200 ms | 200–500 ms | > 500 ms |
| CLS (Cumulative Layout Shift) | How much content moves around while loading | ≤ 0.1 | 0.1–0.25 | > 0.25 |
According to web.dev, a page passes when these targets are met at the 75th percentile of page loads. So even if most visits are fine, a slow quarter of visits can push the result to "needs improvement".
How to use it
- Enter the full URL of the page you want to check.
- Start with the field data. It comes from the Chrome UX Report (CrUX) and reflects the last 28 days at the 75th percentile. web.dev recommends assessing mobile and desktop separately.
- Where configured, lab data from PageSpeed Insights / Lighthouse appears in its own section. Use it to understand where a problem comes from.
- Pick your focus based on which metric is red or amber.
Interpreting the results
Field data tells you what happens; lab data helps explain why. Real users arrive on all sorts of devices and connections, and field data reflects what they actually experience. A lab run is one controlled test, and its numbers can differ from field data. When the two disagree, trust the field data for user experience.
INP can't be measured directly in the lab. It requires real interactions. Lab tools run without a user clicking around, so they typically show indirect indicators, such as long tasks blocking the main thread, instead.
Common causes:
- LCP: Large, uncompressed images; slow server response; a hero image discovered late; render-blocking CSS and JavaScript.
- INP: Heavy JavaScript, third-party tags, event handlers that do too much work on a single click.
- CLS: Images and ad slots without reserved dimensions, banners injected after load, text reflowing when web fonts swap in.
Limitations
- Low-traffic pages may have no CrUX data, and the tool won't guess.
- Field data summarises a 28-day window, so yesterday's fix may not be visible yet.
- A lab run is a single test and can vary slightly each time you run it.
- Google's page experience documentation points out that good results in Core Web Vitals reports don't guarantee your pages will rank at the top.
To watch field data for your whole site every day, alongside a weekly technical crawl, see the SEO site audit.