Learn the metrics
The metrics behind every audit report, explained with the real thresholds and what actually moves them.
Last updated 2026-07-28
- What are Core Web Vitals? Core Web Vitals are three metrics Google uses to measure page experience: LCP, INP and CLS. Here are the exact thresholds and what actually moves them.
- What is Largest Contentful Paint (LCP)? LCP measures when the main content of a page appears. Good is 2.5 seconds or less. Here is what counts as the LCP element and how to make it faster.
- What is Cumulative Layout Shift (CLS)? CLS measures how much a page jumps around while loading. Good is 0.1 or less. Here are the four causes and the fixes that eliminate them for good.
- What is a good Lighthouse score? 90+ is good, 50 to 89 needs improvement, under 50 is poor. Here is what the number actually measures, why it fluctuates, and how much to care.
What are the numbers you need to know?
Five, and they cover most of what an audit report will tell you. Three are Core Web Vitals thresholds measured on real users; two are Lighthouse score bands measured in a lab.
| Metric | Good | Needs improvement | Poor |
|---|---|---|---|
| Largest Contentful Paint | ≤ 2.5s | 2.5–4.0s | > 4.0s |
| Interaction to Next Paint | ≤ 200ms | 200–500ms | > 500ms |
| Cumulative Layout Shift | ≤ 0.1 | 0.1–0.25 | > 0.25 |
| Lighthouse category score | 90–100 | 50–89 | 0–49 |
| Core Web Vitals assessment | All three good at the 75th percentile | Any one in the middle band | Any one poor |
Why do these numbers disagree between tools?
Because they measure two different things: lab data is one simulated visit on one machine, and field data is the 75th percentile of real visits over a rolling 28-day window. They are not supposed to match.
Lighthouse, Audra and Chrome DevTools all report lab data — instant, reproducible, available for any page including staging, but a simulation. Search Console and the field half of PageSpeed Insights report real users on real devices and networks, which is slower to update but is what actually counts for the Core Web Vitals assessment.
When the two disagree, believe the field data and use the lab data to find the cause. Chasing a lab score without confirming the change in field data is the most common way to spend a week on something no user notices. Core Web Vitals explained covers the 75th-percentile rule in detail.
Frequently asked questions
Do I need real user data to check these metrics?
Not to diagnose them. Lab tools like Lighthouse measure LCP and CLS on any page, including staging and password-protected pages, and name the element responsible. You do need field data for the official Core Web Vitals assessment, which requires enough traffic for a CrUX record to exist.
Is a Lighthouse score of 100 worth chasing?
Usually not. Above 90 is good, and the gap between 45 and 85 is where users feel a difference; the gap between 92 and 98 is mostly cosmetic. The scoring curve is log-normal, so each point costs more than the last. See what is a good Lighthouse score.