Site Speed Is Conversion Rate: Core Web Vitals in Plain Language

"Site speed" sounds like something only engineers need to care about — until you translate it into another sentence: visitors who can't wait leave, and every one who leaves is someone you spent money and effort bringing in. Speed isn't a technical metric; it's part of your conversion rate — which is exactly why Google made it a ranking factor and gave it an official name: Core Web Vitals.

This article isn't written for engineers. It's for every site owner who has stared at a report and been baffled by three acronyms: LCP, CLS, INP. Understand these three metrics and you can hold a conversation with any development team — and judge for yourself whether your website is actually healthy.

Three metrics, three kinds of annoyance

The clever thing about Core Web Vitals is that it doesn't measure abstract "fast or slow" — it measures three specific visitor annoyances:

• LCP (Largest Contentful Paint): how long until the main content appears. That stretch of white screen or spinning loader after you tap a link — that's what LCP measures. The standard is within 2.5 seconds; beyond that, the visitor's thumb is already hunting for the back button. The most common culprits are oversized hero images and slow servers.
• CLS (Cumulative Layout Shift): whether the page jumps around. You're about to tap "Confirm," an ad image suddenly loads and shoves the button down, and you tap something else — the rage of that instant is CLS. It measures stability: images without reserved space and layout reflow after fonts load both push this number up.
• INP (Interaction to Next Paint): how long before a tap gets a response. You click a button or open a menu, the page freezes for half a second before moving — that "is this thing broken?" hesitation is INP. Usually it's a page stuffed with too much JavaScript — often years of accumulated tracking snippets and plugins.

The testing tool is free: Google's PageSpeed Insights. Paste in a URL and it gives you a score and a list of what to fix. You don't need to understand every recommendation, but the count of red warnings is enough to tell you whether it's time to talk to someone.

The link between speed and money is more direct than you think

Published studies from every major platform point the same way: each additional second of load time visibly raises bounce rate and lowers conversion; conversely, after speed improvements, conversion and revenue move with it. The exact magnitude varies by industry — we won't quote misleading numbers here — but the direction has never had an exception: no study anywhere concludes "slower is fine."

And the cost of slowness compounds: poor speed means high bounce rate; high bounce rate tells Google the experience is bad; rankings slip, traffic shrinks; so you buy more ads, pouring paid traffic into a leaking bucket. We've seen plenty of marketing budgets that were really just subsidizing a slow website.

Speed is the most honest first impression — visitors start grading you before they've read your first line of copy.

Why some websites are simply born slow

Honestly, many speed problems are decided at birth. Template sites and drag-and-drop builders, in order to "do everything," preload piles of code you'll never use; add uncompressed hero images and layer after layer of plugins and tracking scripts, and every item pushes those three metrics higher. Optimizing after launch is often like putting an overweight car on a diet — possible, but painful, and with a hard ceiling.

This is one reason we insist on hand-building every brand website to fit: nothing loads that isn't used, images are compressed and given reserved layout space at build time, and Core Web Vitals are a factory setting, not an after-sales repair. The same logic matters even more on phones — mobile networks are slower and devices weaker, so speed problems get amplified; we unpack that fully in the mobile experience checklist.

What you can do right now

Three steps. First, test your own site with PageSpeed Insights — both mobile and desktop. Second, compare your scores against the competitors you care about most. Third, if it's a sea of red, deal with images first (usually the cheapest improvement), then assess whether it's a constitutional problem or a maintenance problem. If it's constitutional, face it honestly — that's a redesign signal, not an optimization exercise.