How I got started with ghost

Photo by Matt Duncan / Unsplash

I was looking for a portfolio website for my music, projects and photography in 2017 and after talking to my friends I still remember this phrase:

"Friends don't let friends use Wordpress..."

So James recommended this cool and shiny thing called Ghost instead. I liked the looks and feel of it so lets give it a try! The goal was to have the least maintenance cost and effort and a mobile-friendly clean experience for the users with great editing on the CMS-side. Can't be that hard...

Approach 1: Hosting of Ghost on Uberspace

First I signed up to Uberspace a pretty kick-ass hosting provider that lets you determine your own price. They even had a tutorial on how to install ghost so what could possibly go wrong, right?!

  1. Get the hosting running
  2. Install Ghost
  3. Hook up to custom URL
  4. Keep on Rockin'

So setup was okish, but I failed miserably with the whole SSL and Domain linking which made it ok but not great. Last but not least: Maintenance is a bitch. Meaning updating, backuping etc is not the greatest thing to do and I only broke the install twice by messing with the backend when trying to install new themes.

Approach 2: Gatsby and Ghost hosted via AWS S3

So while the whole experience with self-hosting Ghost was OK, the few euros am month for hosting did not really hurt but triggered my internal Schwabonian (synonym for cheap fuck) which meant an alternative had to be found.

Me trying to save the last €

Using a static website builder from Jamstack with a local CMS, generating the page and then hosting via an AWS-bucket with autoscaling seemed like the cheapest method around (again based on James recommendation). After getting aquainted with static-site generators I found a handy little guide. The idea was simple:

  1. Install Ghost locally and use as a CMS. Edit blog via localhost
  2. Create a static site (all the HTML/CSS) locally
  3. Transfer to S3 Bucket
  4. Hook up to your domain and keep Rockin'

I knew quite a bit about S3 and EC2 via my work in AI so what could possibly go wrong?!

The guide actually used an EC2 instance for hosting Ghost which was unacceptable as these bad-bois cost a lot of moneyz (with a lot being more than 4€/month). The guide was pretty good, but my patience was endless: Next to the job, the motiviation was not to be found to manage a carefully crafted pipeline of build & synch commands, SSL certificates and SSH keys for essentially maintaining a fancy typewriter. I'm still regretfully looking back at this approach as its undoutably the most elegant and also - according to James - increadibly cheap. But maintaining similar pipelines is also his main job so the game was rigged from the beginning.

Approach 3: And now for something completely different

Maybe Ghost and me wasn't meant for each other and there are so many other shiny things like, etc. but my inner nerd could not give fully in to the point and click experience. It felt like giving up...

So after more google searches I found I liked the appeal of this site builder:

One that unifies world-class visual editing with an open, modern developer experience.

The themes were cool, It followed the same principle as the elegant approach above and I liked that I could imagine a little docker starts living somewhere whenever I hit the build button. Its pretty much Approach 2 with a business model made by professionals for professionals.

The website build with Stackbit, great feel and design and constantly improving from their end

The website is fast, looks great and can be customized. Also they host the website and it costs exactly 0€ because they haven't fully figured out their business model! Perfect for planting a little parasitic mini-page and bury it somewhere deep down in their S3-bucket bill!

HOWEVER... the service changed quite a bit even in the short time I was using it. Similar to how Ghost evolved from 2017 to what it is now, they started to introduce SEO-tags but sadly the site I built wasn't backwards compatible. Changing the site to the new format made me mess up what I built because I followed the tutorial they provided but still it was never as before... So in a way again maintenance killed my hopes of a sustainable place where I can grow my content whenever and wherever I feel like.

Approach 4: The downfall

Paying money for a service I can build myself by following free guides on the internet? Never!

While this statement is certainly by me I somehow came to the conclusion that maybe I'm old enough to also follow my brothers advice:

If there are professionals that can build the thing you want to have in a thenth of the time with ten times better quality, then maybe you should pay for it.
my conclusio after three approaches

After only four years of isolated bursts of backend-motivation coming and going, learning about static site builders and implementing a few pipelines I decided to give up and kill it with moneyz.

The real reason I want to start a site like this is to create content, share knowledege and hopefully inspire. Maintaining some hacky backup and update scripts does not necessarily fill me with as much joy as working on blog-posts, so I decided to use my time more wisely. Supporting an open-source community is at least a good cause and maybe helps others in getting an affordable content-creation platform without having to go through too much backend hassle.

The service, functionality and UX of Ghost has improved markedly since 2017 which was one reason why I believe in this project and feel ok giving money. Hopefully the journey will continue that direction. One last quote that might summarize my build-vs-buy journey:

If its not your core competence: license it

Also quote by James btw. I have no original thoughts whatsoever it seems :D

Christoph Götz

Christoph Götz