Core Web Vitals describes loading, responsiveness and visual stability. The three metrics belong together and gain meaning alongside real user and conversion data.
The three metrics and what they measure
Core Web Vitals consists of LCP, INP and CLS. Each metric describes a different part of the experience; view them alongside real user and conversion data.
- LCP (Largest Contentful Paint): how quickly the largest visible content appears. Google's good threshold: 2.5s or less.
- INP (Interaction to Next Paint): how quickly the page visibly responds to interactions. Good: 200ms or less.
- CLS (Cumulative Layout Shift): how much unexpected layout movement occurs. Good: 0.1 or less.
Practical: good INP prevents forms, filters and checkout actions from feeling slow. Then measure in your own funnel whether lower interaction latency produces more completed actions.
The mistakes we see most often
The same causes often recur in performance audits:
- Unoptimised hero images: excessive dimensions, missing srcset or poor compression.
- Third-party scripts blocking render: analytics + chat widget + ads + cookie banner.
- Webfonts without a loading strategy: unnecessary variants and incorrect preloading block text or network capacity.
- Layout shifts from images without dimensions: buttons jump around during load.
What it delivers
After a performance improvement, check three kinds of effect:
- Field data: do LCP, INP and CLS improve at the 75th percentile?
- Behaviour: does abandonment change on mobile landing pages and forms?
- Business KPIs: do enquiries, checkout completion or revenue change with comparable traffic?
Reality check: Lighthouse is a lab test. Combine it with Chrome UX Report field data and analytics; only real visitors show whether improvement is structural.
Conclusion
Performance is not a standalone technical report. Improve LCP, INP and CLS together, monitor regressions and connect changes to real user and business data.