I spent 7 years building my own static site generator CMS.
I added drag and drop, Wordpress-like UI and lot of complex functionality based on studying extensively the existing systems like Jekyll, Hugo, etc. Originally, I started with a Wordpress theme I designed myself. It was great until it wasn't - I tried to add E-Commerce to it and it was the most pathetic architecture for a CMS I've ever worked with. Although I made a successful theme with E-Commerce in it, I refused to use something as fragile. At the time, PHP wasn't great to work with either and I was a Rails guy. So, I looked into Ruby based alternatives and found Middleman and Jekyll. Middleman was quite powerful, but not powerful enough for a travel website where people could sort content by certain criteria, discover new places to travel, etc. Again, I started hitting the limitations of the language itself and felt frustrated. I tried Jekyll, same story. Over the course of these trials, I would work every weekend religiously with meticulous detail. I would spend all my after work hours on this little project of mine.
During the course of these 7 years, many people left, I lost many friends around me because I couldn't trade off the time for them with the time spent this project. I worked on it because I wanted to. It felt like I was going deep into unexplored territory, discovering so many possibilities with just what people would label as a simple CMS. And then, one day, I launched privately. Just to myself. It was a bliss to watch it in action. By then, I had ported it over to Elixir with a custom architecture that could support any business model you want. Data portability and export were one of my core requirements as I had painfully tried to migrate my content from free CDNs more than once after receiving terrabytes of traffic.
Finally, to test my CMS, I implemented it for a client who is still running their entire business on it. They spend barely 1/5th of their competitors' cost and yet are able to serve 3 times the traffic. This was the biggest validation for me, personally. This thing ran by itself for 4 years without any hiccups or maintenance required on the cloud. It was a business unit on auto-pilot for me.
I loved it and it was all because of immensely thanks to the architecture which I consciously disallowed Javascript to not be a part of the whole thing. The frontend was completely separated so deployments to the core/backend wouldn't fail because of some dumb sh*t change Babel made to their config file or Webpack is forcing you to upgrade something. In fact, the JS dependency even for the frontend was quite low and I used Phoenix/LiveView where I could and just static CSS frameworks to make the UI look good and plain old refresh and reload for everything else. This still works well for me.
At its peak performance, the SSG CMS attracted a political party member's attention who was running his own online news site at the time. I got a chance to speak to him over the phone and he was curious how the system worked. I politely told him I couldn't sell him a clone of my system as that would be a betrayal for my client. I valued loyality and would never do something behind his back. He appreciated me for that and got some useful advice to upgrade his stack instead. I had never been more proud.
Finally, as with all good things coming to an end - the contract with my client has terminated and I am now doing Startup School to launch my SSG CMS as a Saas. I am also going to open source the core to enable developers to adopt and reap the benefits of my hard work. I know it is just a simple CMS to many, but I consider it my life's work and I see it as much, much more.
In case you are interested, you can sign up on this page for updates, I plan to launch both the OSS version and Saas version before Q4 this year. Thank you, if you have read so far :)
https://www.trystatic.com