Comparison
Core Web VitalsvsLighthouse score
Core Web Vitals
you were handed three numbers and told the page fails on one of them, at the 75th percentile.
Google's three headline field metrics — largest contentful paint for loading, cumulative layout shift for visual stability, interaction to next paint for responsiveness — each with good and poor thresholds. They are assessed at the 75th percentile of real visits, so an average that looks fine can still fail. They exist because they are the small set of numbers that correlate with users leaving, not because they are the only things worth measuring.
Full entry →Lighthouse score
you moved the number from 62 to 91 and nobody could feel the difference on the actual site.
A weighted composite of several lab metrics, normalised to a 0-100 scale with a deliberately harsh curve. It is useful as a checklist and a rough regression signal, and it is routinely mistaken for the goal itself. Optimising for the score rather than for the underlying metrics is how a team ends up deferring the script that renders the content.
Full entry →