Field Notes: 

The CMS Problem

Connect came first. We'd built it to kill the internal linking grind, and it did the job perfectly: it told us exactly where every missing link should go. But that just shifted the bottleneck along. 

Connect could tell us where two hundred and fifty links needed to sit, but we still had to open two hundred and fifty pages and make every one of those edits by hand. The CMS we were making them in was Apostrophe, a headless CMS that fought us at every turn, so editing 250 different pages was not an afternoon's job. It was months of one.

And to be fair to Apostrophe, it could do the things we needed. It let us build bespoke widgets, with proper custom functionality, and that genuinely mattered to us. Living in it day to day was the problem. The structuring was a fight every single time. The formatting came out inconsistent for reasons you could never quite pin down. Making a minor change caused problems somewhere else - for months, I had to re-line bullet points every time I edited an existing page. Honestly, there was a time when I must have spent days of my life re-lining bullet points. And the visuals were dated to the point where a fresh build looked a decade old before it was even finished.

What flipped it for me had nothing to do with work. I'd knocked together a small site for a creative side project on Google Sites, and the thing was just easy. Plug and play. You drag a block, it lands where you put it and it looks fine. Build columns and embed HTML if you want, and you move on. There's nothing to fight. The catch, naturally, is that there are only a handful of templates, when it comes to widgets, you get what you're given. And the moment you want anything properly bespoke, Google Sites shrugs and tells you no.

So I knew exactly what I wanted, because I'd had both halves in my hands at once: the bespoke muscle of Apostrophe with the get-out-of-the-way simplicity of Google Sites. One tool that did both. I took the question to our IT team. Was there a headless CMS out there that gave us Apostrophe's custom functions and Google Sites' plug-and-play feel?

Their answer was that it might be easier to build our own. So we did, and that became Lennon Tree Pages, the second thing we ever built.

We did it carefully. First we migrated a handful of existing pages across, just to see whether the thing would hold. It held. So we built a one-page site, then a four-page site, then a sixty-page site, and it kept holding. We started standing up standalone international sites in native languages: French, Dutch, Spanish. We took individual products and gave them sites of their own. We built the site you're reading this article on. 

Somewhere in that run it stopped being an experiment and became the way we work. Proof of concept, well and truly earned.

Then we went big. A hundred and seventy page site, followed by the migration that started all this: two hundred and fifty pages lifted off Apostrophe and onto Pages, the CMS we'd grown ourselves.

Here's where it leaves us. We can stand up a fully functional ecommerce site, properly working, in under a day. That capability is worth far more than anything we'd save by renting someone else's CMS. What we were really after was the freedom to move at our own speed, and to change whatever we wanted on the day we decided to.

One day, Lennon Tree Pages will be available for third party usage, if you're interested in using it, but for now it's a part of our suite that remains internal use only. 

- The Gardener. 2026-06-22.