Every performance metric in Calibre is graded against a set of thresholds that classify a measurement as good, needs improvement or poor. These thresholds come from Google’s Core Web Vitals programme and the wider web performance community, and they are the same boundaries Calibre uses to colour metrics across CrUX, Real User Monitoring and Synthetic testing.
This page is the single reference for those values. Each individual metric guide links back here for the full comparison.
How thresholds are assessed#
Core Web Vitals are assessed at the 75th percentile of page loads, split across mobile and desktop. A site passes Google’s assessment when all three Core Web Vitals are rated good at the 75th percentile.
The 75th percentile means three quarters of visits meet the target. Taking Largest Contentful Paint as an example: if 100 people visit your site, LCP needs to be 2.5 seconds or less for 75 of them for the metric to be rated good.
Measuring at the 75th percentile rather than an average means a small number of very fast visits cannot hide a slow experience for everyone else. It reflects what most of your visitors actually experience.
Core Web Vitals thresholds#
These three metrics make up Google’s Core Web Vitals assessment and contribute to its search ranking signal.
| Metric | Good | Needs Improvement | Poor |
|---|---|---|---|
| Largest Contentful Paint (LCP) | < 2.5s | 2.5–4.0s | > 4.0s |
| Interaction to Next Paint (INP) | < 200ms | 200–500ms | > 500ms |
| Cumulative Layout Shift (CLS) | < 0.1 | 0.1–0.25 | > 0.25 |
Supporting metric thresholds#
These metrics are not part of Google’s Core Web Vitals assessment, but they describe important parts of the experience and are useful for diagnosing slow pages and setting performance budgets.
| Metric | Good | Needs Improvement | Poor |
|---|---|---|---|
| First Contentful Paint (FCP) | < 1.8s | 1.8–3.0s | > 3.0s |
| Time to First Byte (TTFB) | < 800ms | 800–1800ms | > 1800ms |
| Total Blocking Time (TBT) | < 200ms | 200–600ms | > 600ms |
Total Blocking Time is a ‘Web Vital’. You can optimise TBT in order to improve first-load interactions, improving Interaction to Next Paint for real users.
How Calibre colours metrics#
Wherever a metric appears in Calibre, it is colour-coded against these thresholds so you can read a result at a glance:
- Good is shown in green.
- Needs improvement is shown in amber.
- Poor is shown in red.
You can see this grading on your CrUX Dashboard, in Real User Monitoring, and on each metric’s detail view. UX Rating builds on the same thresholds to show what percentage of your real user experiences fall into each band.