Core Web Vitals in 2026: What Actually Moves Your Score

Every year a new batch of "Core Web Vitals hacks" makes the rounds, and every year most of them move the needle by a fraction of a percent. After optimizing well over a hundred client sites, the pattern is consistent: a handful of structural decisions account for almost all of the improvement, and everything else is polish.

Largest Contentful Paint starts with the request path

LCP problems are rarely about the hero image itself — they're about how many steps happen before the browser can even start fetching it. Redirect chains, render-blocking CSS, and fonts loaded from a third-party origin are the usual culprits. Preloading the LCP asset and serving it from the same origin as the HTML document typically does more than any amount of image compression.

INP rewards doing less on the main thread

Interaction to Next Paint replaced First Input Delay because it measures the full cost of an interaction, not just the delay before it starts. Long JavaScript tasks — hydration, analytics scripts, and heavy client-side state updates — are what push INP into the red. Breaking up large tasks and deferring non-critical scripts until after the page is interactive consistently brings INP back under 200ms.

Layout shift, meanwhile, is almost always solved the same way: reserve space for images, embeds, and ads before they load, and avoid injecting content above things the user is already looking at. It's unglamorous, but it's the fix that works.

The takeaway: chase the structural fixes first — request path, main-thread work, and reserved layout space — before reaching for micro-optimizations. That's the order we work through on every site we touch.