Core Web Vitals, explained for founders
What the three numbers Google cares about mean for your conversion rate, and the handful of fixes that move them without a rewrite.
- web
- performance
- seo
If you run a product with a public website, Google grades it on three numbers. They are not vanity metrics: on the sites I have measured, moving from "needs improvement" to "good" was worth a 5–12% lift in sign-ups.
The three numbers
- LCP (Largest Contentful Paint) — how long until the main thing on the page shows up. Target: under 2.5 s.
- INP (Interaction to Next Paint) — how quickly the page reacts when someone taps or types. Target: under 200 ms.
- CLS (Cumulative Layout Shift) — how much things jump around while loading. Target: under 0.1.
You can see your real-user scores in Google Search Console under Core Web Vitals. Lab tools like Lighthouse are useful, but field data is what ranks you.
The fixes that usually matter
1. Serve images properly
Oversized hero images are the number one LCP problem. Use a framework image component or a CDN that resizes, set explicit width and height, and preload the hero.
<Image src="/hero.jpg" alt="" width={1200} height={800} priority />2. Load third-party scripts late
Chat widgets, analytics and tag managers are the usual INP culprits. Load them after the page is interactive, and ask whether you need all of them.
3. Reserve space for everything that loads later
Ads, embeds, cookie banners and fonts cause layout shift. Give containers a fixed aspect ratio, and use font-display: swap with a matched fallback font.
A page that scores 95 in Lighthouse but ships a 400 KB chat widget on mobile will still fail INP in the field. Measure on a cheap Android phone.
When to call someone
If your scores are red and your site runs on a page builder or a legacy theme, a one-week audit usually finds the three changes that fix 80% of it. A rewrite is almost never the first step.
Marco Restif
Software engineer · AI & full-stack. Writing about what I build.
Choosing a tech stack for your MVP in 2026