Beta

This is a new service and pages are being tested and improved.

I’ve been a content designer in Government Project Delivery for a year now. It took time to work out what the role would look like, what does content design in project delivery even mean?

For most of the last year, I’ve been the website guy. Need something published? Ask Carwyn. Page not working? Ask Carwyn. That’s fine. I like the website. But what’s great is that increasingly I’m involved earlier. Before something gets written. Before decisions get made about format or structure.

Some of you have sent me things to publish, asked me to look at a draft, or had me explain why I’ve restructured something you sent over. Some of you have probably had enough of me asking questions, wanting to know how something sits in the bigger picture or comment reviews on work. But asking all these questions really is a lot of the job.

The Teal Book

The Teal Book is a good example of how we can do things a little differently. It’s the definitive guide for project delivery in government. You know this. Many of you shaped it.

It could have been a PDF, published once and left alone. That’s how most guidance works. Instead, it launched in April 2025 as a digital product. Guidance spread across pages where signed-in members can comment on any line, at any time.

Over 250 organisations have given feedback. We’re continuously reviewing and publishing changes in response contents. We are going through the latest comments now to shape version 3.0.

That feedback model matters. People spot things we missed. They tell us when something doesn’t match how things actually work. They ask questions that show where we weren’t clear enough.

Project delivery people already think this way about portfolios, programmes and projects. Test your assumptions. Learn as you go. Adapt. Content design applies the same logic to guidance and communications.

What the numbers say

In August 2024, our site appeared in Google search results about 52,000 times per week. Last week, that number was over 250,000. This January we had 14% more page views from 57% more users than January 2025.

That came from a few things: time since going public last March, thinking about what people actually search for and using those keywords, structuring content in a way that answers questions, writing titles that describe what users will get and writing good descriptions for pages.

Now we’re looking at AI and how people use Copilot, Gemini, ChatGPT and others. If our content isn’t structured in ways those tools understand, it won’t get used. So we’re making sure AI tools understand what content is about and how it answers people’s questions. We’re carefully opening up to certain bots, to make sure authoritative government guidance is what AI learns from.

There’s not much guidance out there on how to do this well. We’re figuring it out.

Two products I want you to know about

I’ve recently published a content type guide and a design system.

The content type guide helps you choose the right format before you start writing. What does the user need to do? What format serves that? It’s about starting with the outcome, not the format you fancy using.

When content types are consistent, people know what to expect. A guide is a guide. A standard is a standard. Teams can see what exists and build on it instead of creating their own version of the same thing.

The design system documents how we present things on the website. Every component. When to use it. When not to. Why. This stuff used to live in my head. Now it’s written down. Anyone can see the decisions and challenge them. That’s the point.

The bit I care about most

The thing that frustrates me is when you ask why something is done a certain way and the answer is “that’s how we’ve always done it”.

Content design is basically just refusing to accept that. Does this work? How do we know? What should we try instead?

The design system and content type guide are ways of working as much as they are products. They’re how we make decisions visible, so anyone can question them and we can keep improving.

That’s the approach I hope sticks. Keep asking if things work. Keep changing them when they don’t.

Author

Carwyn Williams

National Infrastructure and Service Transformation Authority

London
Back to top