In this article
In October 2026 we rebuilt the CL42ITY website and launched it. It looked the way we wanted, but Google's PageSpeed test gave the mobile version 61 out of 100. Within a few days it scored 97 to 100. This is what we found and what we changed, in plain English, because the same lessons apply to most small business websites.
Why the score matters
Google uses real-visitor speed, called Core Web Vitals, as one of many signals when it ranks pages. More importantly, slow pages lose visitors. People on phones give up quickly, and every second of waiting costs enquiries. A fast site is simply a better shop window.
Lesson one: the test isn't your phone
Our first surprise was that the site felt quick on our own phones and laptops, yet scored badly. Google's test runs on servers with no graphics chip and simulates a slow 4G connection. Anything that leans on the graphics chip, such as heavy blur effects or 3D, becomes very slow in that test. We measured the site the same way Google does, which showed us the real problems straight away.
Lesson two: what loads before the first paint counts most
Google's estimate of how fast your page appears includes every important file the browser has to fetch before showing anything. Our website builder at the time loaded about 400KB of its own code early, plus two web fonts from Google's servers. We moved the site to plain HTML on Cloudflare, built our brand font into the page and loaded the second font only when the reader reached the section that uses it.
Lesson three: don't let one feature block the page
Our homepage has an interactive glass effect you can break. On computers without a graphics chip, setting it up froze the page for over a second. We changed it so it only starts when someone moves the mouse or touches the screen, checks for a real graphics chip first and builds its parts in the background. Visitors still get the effect; the page never waits for it.
Lesson four: images and layout shifts
Every image now has its width and height set, so the page doesn't jump as pictures load, and phones only download the size they need. Long grids of project photos are only drawn when you scroll near them. That one change took our work page from 78 to 100 on mobile.
The results
- Mobile PageSpeed: 61 before, 97 to 100 after.
- Desktop PageSpeed: 64 in a third-party test, 98 to 100 after.
- Accessibility, best practices and SEO: 100 on every page.
- Real visitors: pages load in about 0.6 seconds on the main content.
What this means for your website
You don't need to rebuild everything to get most of the benefit. Start by testing your homepage at pagespeed.web.dev on the mobile tab, then look at three things: the size of your images, how many fonts and scripts load at the start, and any heavy effects or sliders at the top of the page. If your site is on a builder, ask what you can switch off.
If you'd like us to look at yours, our website redesign service starts with a speed and SEO check, and every website we build is tested this way before it goes live.
