The Craft dimension of the rubric includes measured load performance. This page states exactly which thresholds are used and where they come from, so that a Craft score can be checked rather than taken on trust.
The three metrics
Core Web Vitals are three metrics Google publishes to describe real-world user experience. Each is assessed as good, needs improvement, or poor.
| Metric | Measures | Good | Needs improvement | Poor |
|---|---|---|---|---|
| LCP | Loading. Time until the largest content element renders | 2.5s or less | 2.5s to 4s | Over 4s |
| INP | Responsiveness. Latency across all interactions in a visit | 200ms or less | 200ms to 500ms | Over 500ms |
| CLS | Visual stability. How much layout shifts unexpectedly | 0.1 or less | 0.1 to 0.25 | Over 0.25 |
How the assessment works
Google evaluates each metric at the 75th percentile of real user data, drawn from the Chrome User Experience Report. That means at least 75 percent of visits must meet the good threshold for a page to pass that metric. A page passes Core Web Vitals overall only when all three metrics are good at that percentile.
The 75th percentile is a deliberate choice. It captures the experience of most visitors rather than the median, which would ignore the slower half, without letting a handful of outlier visits on poor connections define the result.
INP replaced First Input Delay in March 2024. FID measured only the delay before the browser began processing the first interaction. INP measures responsiveness across every interaction in the visit, which makes it a more complete and generally less forgiving measurement.
Field data and lab data are not the same
Core Web Vitals as Google assesses them are field data: measurements from actual visits, collected through the Chrome User Experience Report. Tools like Lighthouse produce lab data, a simulated load under fixed conditions.
Lab data is useful for diagnosis because it is repeatable and tells you what to fix. It is not what Google assesses. A site can score well in Lighthouse and still fail Core Web Vitals if real visitors are on slower devices or connections than the simulation assumes.
How this is used in a teardown
Craft is scored on more than performance. Type scale consistency, line length, the mobile reading experience and semantic markup all count. But performance is the part that can be measured identically across every subject, so it anchors the dimension.
Where field data is unavailable, which is common for smaller publications that do not meet the reporting threshold in the Chrome User Experience Report, the teardown says so and relies on lab measurement, flagged as such. A lab number is never presented as though it were a field number.
Sources
- web.dev, How the Core Web Vitals metrics thresholds were defined
- Google Search Console Help, Core Web Vitals report
Thresholds verified August 2026. Google has revised Core Web Vitals before, including replacing FID with INP, so this page records the version of the metrics each teardown was scored against.