I’m trying to understand why most Next.js websites — including vercel.com — tend to have poor LCP (Largest Contentful Paint) scores on mobile in Core Web Vitals. While desktop performance is usually solid, mobile scores consistently fall short (see screenshots for reference).
Vercel website mobile score
LCP is Heading
I’ve done some research and tested multiple websites built with Next.js. In every case, mobile LCP scores were poor — particularly in PageSpeed Insights. Interestingly, Lighthouse (run locally or in dev tools) doesn’t show such severe results. I always test using an anonymous window, so caching or extensions shouldn’t be a factor.
How is it possible for a heading element to have an LCP time of 9 seconds?
Could this be caused by Google’s default mobile throttling during testing? And should I even care this much about Core Web Vitals on mobile?
Is it realistically possible to achieve a 90+ score on mobile using Next.js?
I’m especially concerned because Core Web Vitals affect search rankings, and mobile performance is often the primary benchmark.
EDIT:
Since Lighthouse reports are more forgiving than PageSpeed Insights, what’s the most reliable way to measure real-world Core Web Vitals performance? Any advice on how to test or optimize more effectively?
