Site speed in 2026 — what is still hurting your LCP and what to do about it

Core Web Vitals have been around for a while, but the number of sites with poor LCP (Largest Contentful Paint) scores remains surprisingly high. According to Chrome's CrUX field data, only around 55% of sites globally achieve a 'Good' LCP below 2.5 seconds. For Serbian business sites, that figure is lower.

The most common culprit in 2026 is not the same as in 2022. Images were the top issue — and they still contribute — but WebP conversion and lazy loading have become standard. The bigger problems now are font rendering blockers, third-party scripts (chat widgets, analytics, ad tags), and slow server response times on shared hosting that is not scaled for traffic.

For Next.js sites — we use it for Feather Studio and for client projects — a specific problem is JavaScript hydration on mobile. A large JavaScript bundle must be downloaded, parsed, and executed before the page becomes interactive. React Server Components (experimental two years ago, now a stable part of Next.js) dramatically reduce the JS sent to the client.

A practical test you can run right now: open PageSpeed Insights, enter your site URL, and look at the filmstrip screenshot. Which element was the last to render in the LCP score? That is your target. It is usually the hero image, hero text, or a banner — each with a different fix.

On the server side, improving Time to First Byte (TTFB) below 800ms is a prerequisite for good LCP. Shared hosting abroad with a TTFB of 1.2 seconds cannot produce a good LCP without a CDN. Vercel, Cloudflare Pages, or a decent VPS in Frankfurt makes a significant difference for Serbian users.

If your PageSpeed Insights mobile score is below 70 and you are not sure how to improve it, see our web development services — we do both performance audits and optimisation of existing sites.