Core Web Vitals in 2025: What UK Businesses Need to Know

First Input Delay measured only the delay before the browser could begin responding to the first interaction on a page. The problem with FID is that it measured the wait, not the response. A page could have a great FID score while feeling sluggish every time a user tapped a button or expanded an accordion.

INP measures the full interaction latency — from the moment a user interacts with any element on the page (a click, a tap, a key press) to the moment the next frame is visually updated. It captures the worst interaction during the entire session and uses that as the score. This is a fundamentally harder metric to pass.

Google's thresholds: INP below 200ms is "Good", 200–500ms "Needs Improvement", above 500ms is "Poor". Most UK business websites built before 2023 score in the Poor range without knowing it.

The root cause in the vast majority of sites we audit is the same: too much JavaScript executing on the main thread. Modern sites built with React, Next.js, or other heavy frameworks often ship hundreds of kilobytes of JavaScript that blocks the browser's ability to respond to user input.

Tag managers are a second major contributor. A site with Google Tag Manager loading 12 third-party scripts — analytics, chat, heatmaps, remarketing pixels — creates unpredictable main thread contention. Each script competes for processing time with your actual UI code.

Event handlers attached to elements are a third area. If a click handler triggers a synchronous operation — a filter, a calculation, a DOM manipulation — the browser cannot paint the result until that operation completes. 600ms of synchronous JavaScript means a 600ms INP.

Start with PageSpeed Insights (pagespeed.web.dev). Enter your URL and select the mobile report — mobile scores are the primary signal Google uses for ranking. Look for the INP metric under Field Data if your site has sufficient traffic to appear in the Chrome User Experience Report (CrUX).

For lab-based testing, Chrome DevTools' Performance panel is your best tool. Record a session while interacting with the page — clicking buttons, submitting forms, opening dropdowns. Look for red bars in the interaction track and long tasks (tasks over 50ms shown with red triangles) on the main thread.

The "Interactions" section of a Performance recording shows you precisely which interactions are slow and exactly which JavaScript executed during them. This is where diagnosis happens.

Break up long tasks. Any JavaScript task running for more than 50ms risks blocking input. Use scheduler.yield() (or setTimeout with a 0ms delay as a fallback) to yield back to the browser between chunks of work. React 18's Concurrent Mode and startTransition are purpose-built for this pattern.

Move non-critical work off the main thread. Web Workers can handle data transformations, search indexing, and other CPU-heavy tasks without blocking UI responsiveness. This is especially valuable for filter operations in e-commerce catalogue pages.

Audit your tag manager payload. Remove every third-party script you cannot attribute to direct business value. Load analytics asynchronously and after the page becomes interactive. A 30% reduction in third-party script size typically produces a measurable INP improvement.

For React applications specifically: avoid synchronous state updates in event handlers that trigger heavy re-renders. Wrap expensive state updates in startTransition to tell React they are non-urgent, allowing it to keep the UI responsive while the update processes.

While INP is the newest challenge, Largest Contentful Paint (LCP) and Cumulative Layout Shift (CLS) remain in the Core Web Vitals set and are equally weighted in Google's page experience signal.

LCP (target: under 2.5 seconds) is most commonly hurt by unoptimised hero images, render-blocking stylesheets, and slow server response times. Preloading your LCP image with a rel="preload" link tag, serving modern formats (WebP, AVIF), and ensuring your server responds in under 200ms will fix 80% of LCP problems.

CLS (target: under 0.1) is caused by content that shifts after initial render — images without dimensions, late-loading web fonts causing text reflow, dynamically injected banners. Explicit width and height attributes on images and font-display: swap with preloaded fonts resolve most CLS issues.

Google has confirmed that Core Web Vitals are a ranking factor, used as a tiebreaker between pages of comparable relevance. For competitive UK search terms — business insurance, property management software, B2B SaaS — this tiebreaker is often the difference between page one and page two.

More practically: better INP means lower bounce rates and higher form completion rates. Users who tap a button and see it respond immediately stay; users who tap a button and wait half a second have already learned that your site is slow. Fixing INP on a B2B lead generation site consistently improves conversion rates by 8–15%.