The Page Experience Update Wasn't a One-Time Event
When Google rolled out Core Web Vitals as ranking signals, most SEO coverage treated it like a one-time sprint: fix your LCP, patch your CLS, and move on. That framing was wrong, and it's costing websites real traffic every day.
Core Web Vitals are measured continuously in the field — meaning your score changes as your content, third-party scripts, and user behavior evolve. A site that passed CWV checks in 2023 can be failing them today because a new chat widget got installed, an image format changed, or your hosting provider throttled CPU.
The Three Signals (And What Actually Breaks Them)
Largest Contentful Paint (LCP) measures how long it takes for the biggest visible element to load. The threshold is 2.5 seconds. The most common culprits we see in audits:
- Hero images served as PNG instead of WebP or AVIF
- Render-blocking CSS loaded synchronously in
- Slow server response times above 600ms TTFB
- Lazy-loading applied to the LCP element itself (this is a common mistake)
Cumulative Layout Shift (CLS) measures visual stability — how much the page jumps around as it loads. Score must be below 0.1. Typical killers:
- Images and iframes without explicit width/height attributes
- Dynamically injected banners, cookie notices, and chat widgets
- Font swaps on custom web fonts without
font-display: optional
Interaction to Next Paint (INP) replaced FID in 2024 and is the hardest to fix. It measures the worst input delay across a session. Common issues:
- Long JavaScript tasks blocking the main thread
- Third-party scripts running on every interaction
- React/SvelteKit hydration overhead on large components
How to Find What's Breaking Your Site
The fastest way to know your real-world CWV scores (not the lab simulation) is to check the Chrome User Experience Report (CrUX) via Google Search Console under Core Web Vitals → Open Report. Lab tools like Lighthouse show you what *could* happen; CrUX shows what *is* happening with real users.
If your field data is poor but your lab scores look fine, the gap is almost always caused by:
1. A third-party script that loads after the page appears complete
2. Mobile users on slower connections (CrUX is device-separated)
3. A specific page template that gets high traffic but wasn't tested
The Rankings Connection Is Real
A 2025 study analyzing 500,000 page rankings found that pages in the "Good" CWV tier ranked, on average, 15% higher in competitive queries than equivalent pages in the "Needs Improvement" tier — even when domain authority was held constant.
The effect is strongest in verticals where Google sees high page-switching rates: local services, e-commerce, and news. If users bounce before your page renders, Google notices.
What to Fix First
Run an audit, look at your LCP image, and serve it pre-loaded with a tag. That single change fixes LCP on most marketing sites within 15 minutes and delivers measurable ranking improvement within 28 days (Google's CrUX data update cycle).
After that: compress and convert images, remove render-blocking resources, and audit your third-party scripts ruthlessly. If a script is not generating measurable revenue, it should not be on your site.