Setting a New Standard for ISO.org

Written by Harry Roberts on CSS Wizardry.

ISO is the standard in standards. So imagine my delight when they got in touch asking me to make ISO.org faster for their millions of monthly visitors!

I love a spec, I love a standard, so working with the International Organization for Standardization is a bit of a dream come true! As a Brit, ISO 8601 and ISO 3103:2019 are particularly close to my heart, so to work with the folk that wrote them…? It’s a yes from me.

Conversations started in October 2024 with a view to kicking off the project in spring 2025. This gave us ample time to design the perfect engagement and be very well prepared into the run-up. The brief was simple: improve the user experience. The focus wasn’t on SEO or revenue, but simply on improving the user experience. After all, improve UX, and all else follows. It’s a wise move!

We decided to benchmark against Core Web Vitals as they’re arguably the most sensible place to start for a fledgling project. Their biggest sticking points were Largest Contentful Paint and Cumulative Layout Shift. Interaction to Next Paint wasn’t much of a concern for them as they’d made smart decisions in the past and not gone all-in on JavaScript.

The project was to be incredibly tactical—a quick-fire round of specific improvements built out into a backlog that the team could pick up as and when they were ready. What was nice about this approach, and not running a big reveal-style project, is that we were able to realise performance improvements while the project was in flight. Not only was this highly motivating, it demonstrated immediate value. It was definitely the right thing to do!

And how did it go? It went very well:

You can see in the graphs above that our first performance-facing deployments began in the last week of March. Taking the CrUX data for then versus the time of writing (July 2025), the headline results are:

• LCP was improved by 800ms, from 2.8s down to 2s.
  • A 29% improvement!
• INP was improved by 7ms, from 85ms down to 78ms.
  • It’s an 8% improvement, but at this scale it simply doesn’t count.
• CLS was improved by 0.08, from 0.18 down to 0.01.
  • This represents a 94% improvement!

The team—and I!—are understandably very very happy with these results. The project was a short one, and I have ideas of how we can take these numbers even further with a little bit more of a push, but for the first time ever, ISO.org is all green.

As is customary, I handed over a full Trello backlog (half of which was already live), and an Executive Summary document for the non-technical stakeholders and sponsors.

It was about as targeted as a web performance audit can get: in and out, job done.

If you need some of the same, get in touch.

---

Next up:

Cache in the cloud—fixing caching at Cloudinary

View all case studies…