Every site we audit starts the same way: open Lighthouse, run the report, and watch the founder’s face when LCP shows up red. The reflex is to call it a developer problem — slow images, heavy JavaScript, lazy hosting. That framing misses what’s actually broken.
Core Web Vitals are a proxy for how it feels to use your site on the device your customer is actually holding. LCP measures the time until the page looks usable. INP measures the time between a tap and a visible response. CLS measures how often layout shifts cause a mis-tap. Each one maps to a moment of friction that costs you a transaction.
What slow vitals actually cost
A 1-second LCP improvement on mobile is associated with measurable lifts in conversion across e-commerce, lead gen, and SaaS. Not because Google rewards it (though it does), but because the customer never had to decide whether your site was worth waiting for.
The same is true for INP. A button that takes 400ms to respond gets tapped twice. A booking flow that shifts layout mid-tap routes the user to the wrong CTA. These aren’t edge cases — they’re the median experience on a mid-range Android.
Where we’d start
In every engagement we open with a vitals audit before we touch the design. The fastest wins are almost always:
- Server-side rendering the first paint
- Cutting unused JavaScript (every kilobyte is parsed before the page is interactive)
- Self-hosting fonts with
font-display: swap - Reserving space for images and embeds so CLS doesn’t spike
The boring fixes outperform the clever ones. We’ve seen sites pull LCP from 4.2s to 0.9s without changing a single pixel of the design — just by making the stack honest about what’s on the page.
If your vitals are red, the cost isn’t a worse ranking. It’s the customers who close the tab before they ever met your product.