I Was Doing Fine Building Websites. So Why Did I Get Into This Mess?

I was doing fine. I built websites for small businesses and for people starting something of their own, I had more requests than hours, and the work paid. If you had asked me a few years ago whether I planned to make my own site builder, seventeen WordPress themes and a family of plugins, I would have laughed and gone back to the client in front of me.

So this is the honest answer to a question I now ask myself on some evenings: why did I get into this mess?

The work was fine. The sites were not

Most of my projects did not start from nothing. A client would come to me with a site that already existed, built by someone else, and a problem they could not solve on their own. Sometimes the problem was small. Often it was the same few problems wearing different clothes.

The first one was ownership. A surprising number of clients did not actually control their own website. The domain was registered in the name of a previous designer who had stopped answering emails. The hosting account belonged to an agency that charged a monthly fee for “maintenance” nobody could describe. The site lived on a hosted builder that had no way to take it anywhere else. The client paid for all of it and owned almost none of it.

The second was construction. Underneath a decent-looking home page there was often a mess: a theme loaded with options nobody used, a page builder that wrapped every paragraph in five layers of code, a dozen plugins doing jobs that overlapped. The site looked finished. It was slow, fragile and painful to change.

The third arrived more recently. People started asking AI assistants the questions they used to type into a search box. Google began showing AI answers above the results in the United States in May 2024, and OpenAI launched search inside ChatGPT at the end of October that year. Most of the sites I was asked to look at had never been built with that in mind. They gave machines very little to work with.

The fourth was neglect. Sites that had been fine at launch were running outdated versions of WordPress, a theme whose developer had moved on, and plugins that had not been touched in years. Nobody was looking after them, because looking after them had never been part of the deal.

Fixing it by hand stopped making sense

For a long time I fixed these problems by hand, site by site. I moved domains back into the client’s name. I rebuilt sites on clean foundations. I added the structured information that search engines and AI tools read. I set up updates and checked them.

It worked, but it did not scale, and the worst part was the repetition. Every migration meant working inside someone else’s setup, with their logins, their hosting panel and their odd decisions. Every rebuild meant taking a popular theme and spending hours removing what I did not need and adding what I did. I was solving the same problems over and over, and charging clients for my time to do it.

At some point the obvious thought arrived: if I keep doing the same steps, I should turn the steps into tools.

From fixes to tools

The first thing I systematised was moving sites. Getting a client’s site out of wherever it was trapped and into a clean WordPress installation they owned was the job I did most often, so I built a migration process for it. Out of that work came two products. NoTiesBuilder lets people build a site that is theirs from the first day, and FreeYourWeb helps people get an existing site out of a platform that will not let it go. NoTiesBuilder is, honestly, a by-product of the migration work: once you know exactly how to build a clean, portable site at the end of a migration, you can offer that same result to people who are starting from scratch.

Then came the themes. I was tired of customising the same few popular themes for every client, so I started building my own, each for a specific kind of site. Then the plugins, because WordPress has rules about what a free theme may contain, and some of what my themes needed had to live in a plugin instead. I explain both in their own articles.

All of this now sits under The Web Architecture, the name I gave to the whole family of products.

The four rules

When I look back at every fix I made, they all come down to four rules. I now use them as the test for everything I build, and they are the four themes of this site.

The site is the client’s. The domain and the hosting are in the owner’s name, they have full administrator access, and they can leave and work with someone else without asking anyone for anything.

It is well built. Clear structure, fast pages, sensible code underneath, nothing it does not need.

It is ready for AI search. It describes the business and its pages in a way search engines and AI assistants can read, and it answers questions clearly.

It is kept current. WordPress, the theme and the plugins are up to date, and nothing abandoned is left installed.

These are also the four standards behind the Web Architect Certified programme, which I set up so other builders who work this way can say so. More on that in its own place.

So, was it worth it?

I still ask myself that on the evenings when a theme fails a review for the third time. The answer keeps coming out the same. The problems I kept fixing were not rare. They are how a large part of the web is built and sold, and most of the people affected do not know it until they try to leave, change something or find themselves missing from the answers people now read.

This site is where I explain, in my own words, why each piece of what I build exists. Start with whichever of the four rules matters most to you.