Core Web Vitals Aren't an SEO Trick. They're Your Revenue

Last quarter we audited an e-commerce site that "passed" the PageSpeed test on desktop and failed on mobile. The owner was convinced Core Web Vitals were some Google bureaucracy you tick off so your ranking doesn't drop. When we showed him that 68% of his traffic came from mobile, and that LCP on that mobile was 4.8 seconds, his tone changed. Because this was never a story about Google. It's a story about buyers who leave before the page even shows them anything.

That's the core misunderstanding around Core Web Vitals. People treat them as an SEO checklist item, something a developer clears to turn the score green. Then they wonder why a redesign that "passed every test" doesn't sell more. The answer is simple: speed doesn't affect ranking as much as it affects whether a human sticks around at all.

What Google actually measures

Three metrics, and they're worth knowing by heart because people mix them up constantly:

• LCP (Largest Contentful Paint) measures when the largest visible element appears, usually the hero image or headline. Good is under 2.5 seconds.
• INP (Interaction to Next Paint) measures how much the interface lags when a user clicks or types. Good is under 200 milliseconds. This is an important change: INP officially replaced the old FID in March 2024 and is far stricter, because it measures the full lifecycle of an interaction, not just the first tap.
• CLS (Cumulative Layout Shift) measures how much your content "jumps" while loading. Good is under 0.1.

If you last looked at your metrics before FID was swapped for INP, there's a good chance your score is no longer what you think it is. INP catches what the user actually feels: a laptop that freezes for a second after you click "Add to cart".

The numbers that change the argument

This is where I usually lose the person who thinks this is about aesthetics. Deloitte ran a study with Google called "Milliseconds Make Millions" across 37 brands in retail, travel, luxury and lead generation. They isolated speed as the variable and tracked what happened to business outcomes.

The result: a mere 0.1 second faster load on mobile lifted retail conversion rates by roughly 8%, and travel by roughly 10%. Retail spend went up nearly 10%. That's not marginal. That's the difference between a quarter that hit plan and one that missed it.

Think about the mechanism. A slow site doesn't just lose the sale at checkout. It loses it at every step of the funnel. Higher bounce on the landing page, fewer product views, more abandoned carts. Every extra second of load is a tax you pay at each step, and the client rarely sees it in analytics, because those people simply never enter the "users who saw the product page" statistic.

Why agencies that ignore this hurt their clients

Let me be direct here. Plenty of agencies ship a gorgeous Figma design, stuff it with five third-party scripts for analytics and chat, drop in a three-megabyte hero video, and then wonder why INP is red. They tell the client "we passed the design review" and invoice. But a design that kills performance isn't a finished design, it's a half-built product.

On one project for a travel brand we ran into exactly this. The legacy site had a homepage carousel that loaded six full-resolution images before anything else could start. LCP was over four seconds. We fixed it without adding a single new design element: lazy loading for everything except the first image, modern image formats, and removing two scripts nobody used. LCP dropped below two seconds. We didn't touch the look. We touched what the user feels.

Where the fastest wins are

If you're on a limited budget, don't attack everything at once. The order that usually gives us the most for the least:

1. Images. The biggest LCP culprit. Modern formats, correct dimensions, lazy loading below the fold.
2. Third-party scripts. Analytics, chat widgets, cookie banners. Measure what each one costs and drop what doesn't earn its keep.
3. JavaScript that blocks interaction. This is where INP breaks. Break up long tasks, defer whatever isn't needed for the first render.

You don't need a perfect 100 on PageSpeed. You need green on the field data from real users, because that's what turns into revenue.

Related:

• We Stopped Chasing Design Trends. Here's Why Your Site Wins
• Digital Accessibility: Why It Matters for Your Business

If you're not sure where your site is losing money on speed, we measure it on real users, not on a synthetic test that paints everything rosy. Contact us and we'll show you where the lever is.