Site Speed and Core Web Vitals: Why a Slow Site Raises CPC and Cuts ROAS

There's a number every e-commerce owner can check in 30 seconds that directly affects how much each Google Ads click costs and how many buyers quietly leave before the page even loads. That number is site speed, and Google measures it through Core Web Vitals.

Most stores run WooCommerce, Shopify or Magento on shared hosting, with 15 plugins and a theme that loads 3 MB of JavaScript before the first product is visible. The result: a mobile LCP of 5–8 seconds. And 53% of mobile visitors abandon a page that takes longer than 3 seconds to load. You pay for the click, the buyer leaves before seeing anything.

01 — What Are Core Web Vitals

They are three metrics Google uses to measure the real user experience. Since March 2024, INP replaced the old FID.

Metric	What it measures	Good result
LCP (Largest Contentful Paint)	When the largest element loads (usually the hero image)	< 2.5 s
INP (Interaction to Next Paint)	How fast the site responds to a click/tap	< 200 ms
CLS (Cumulative Layout Shift)	How much content „jumps" while loading	< 0.1

Core Web Vitals are to Google what an expiry date is to a grocery store: an invisible threshold below which no one punishes you publicly, but they quietly take you off the shelf.

02 — Why Speed Affects the Cost of a Google Ads Click

This is the part most often overlooked. Google Ads doesn't charge by bid alone — it charges by Quality Score, one of whose three components is landing page experience. A slow site = lower Quality Score = higher CPC for the same position. A competitor with a faster site pays less for the same click.

The chain is simple: slow site → poor LCP → lower landing page experience → lower Quality Score → higher CPC → fewer clicks for the same budget → lower ROAS. Speed isn't a „nice to have", it's a multiplier on your entire ad spend.

03 — Why Classic Optimization Has a Ceiling

A caching plugin, image compression, lazy-load — all of it helps, but hits a wall. The reason is architecture: traditional WooCommerce/Magento renders the page on the server on every request, pulls from the database, runs PHP, and only then sends HTML. On shared hosting that's 800 ms+ just to the first byte (TTFB).

• Caching plugin — helps, until the next theme update breaks it.
• Image optimization — mandatory, but doesn't fix JS bloat.
• Better hosting — buys you a second, doesn't fix the render model.
• Deleting plugins — helps until you need the functionality you deleted.

Each of these is patching. The ceiling exists because you're still rendering a heavy page on a weak server in real time.

04 — Headless: Why It Breaks the Ceiling

The headless approach separates the backend (your WooCommerce/Shopify, where products, orders and stock live) from the storefront (the frontend the buyer sees). We rebuild the storefront on a modern stack (Next.js / React) served as static, pre-rendered HTML from a CDN — geographically close to the user, with no waiting on PHP and the database.

Specifically what changes for Core Web Vitals:

• LCP drops because HTML is served from an edge CDN, and images are in WebP/AVIF and properly sized.
• INP improves because we ship only the necessary JavaScript, not 3 MB of plugin code.
• CLS goes toward zero because element dimensions are reserved up front.

Important: the backend stays. You don't change platform, don't migrate products, don't touch the admin. Only what the buyer sees and what Google measures changes.

05 — How to Measure Where You Are Now

1. Open pagespeed.web.dev and enter the URL of your busiest product page.
2. Look at the mobile result, not desktop — that's where 70%+ of traffic is and where Google measures.
3. Distinguish Lab data (Lighthouse simulation) from Field data (CrUX, real users). Field is what Google ranks you by.
4. If mobile LCP is over 2.5 s or the overall result is red — caching won't be enough.

Speed is the same principle as the feed: it's not marketing, it's data and delivery engineering. A bad feed Google doesn't see; a slow site the buyer never waits for.

---

Follow the blog for the next part: how we measure the impact of speed on ROAS in GA4, and which images hurt LCP the most.