Why We Wrote a Blog
We just shipped a new version of our blog. It’s a Rails app. It’s not some fancy-pants, johnny-come-lately static site generator. Why on Earth would we do such a thing?
Well, it turns out there are several reasons.
We write Rails apps everyday. We write them for our clients, we write them for our own internal use and even for prototyping. There’s a lot of value in the convention that comes with working on a Rails app.
Rails is crazy powerful. It’s easy to forget. When you have a well defined problem and want to solve it, you can move super fast with Rails. This is it’s sweet spot: rendering HTML pages and getting them to the browser.
Static sites are useful. I won’t disagree, but Rails also has a crazy powerful toolchain built in that lots of projects are attempting to replicate. I know I can use tilt, sprockets, etc. outside of Rails, but why? Lately I’ve adopted a new phrase: “give up and use Rails”. It’s really nice to stop fighting against some preprocessor library and just build things.
When the blog was hosted on Tumblr, editing the theme was basically a nightmare. Yes, I know about Fumblr. And Thumblr. And Thimblr. And all the other half-baked attempts at replicating the Tumblr API locally. It’s just too hard.
Our design team works in Rails apps everyday. In less than a week, we went from a disconnected, difficult to change place to host content to something that uses all our brand styles out of the box.
We write our posts in Markdown and edit them on GitHub. We like the pull request work flow for notifying others to proofread and suggest edits. Sometimes we’ll suggest something that might have been missed.
Once we’re finished editing, Joel posts the article to the blog on his schedule. Tumblr makes that really hard, since Joel couldn’t log in and post the article as the author. We’re just using ActiveAdmin, so it’s stupid simple.
It’s now really easy for anyone on the team to suggest a change. It’s open source, after all. Make a change, submit a pull request. Bam!
We’re no longer beholden to the Tumblr API, we own all of our data now. We like to mix Google Analytics data with author data, which is notoriously missing from the Tumblr API. Now, we can attach that and any other arbitrary data we want to! Did this post generate a lead? Let’s track that!
We have a ton of great content on the blog and podcast. I want to learn how to make it more useful to move around and see related posts. I can do that now. Not to mention, we can roll out and define our own search functionality to increase that engagement. I could even do some A/B testing with something like Vanity, too!
We had a lot of content all over the place. We changed our domain name last year, so we added Rack::Rewrite to handle the old URLs and update them to the new locations. That’s a big help for Google juice.
You can see my half baked redirect database on GitHub in all it’s glory.
We need more, but we’ve got some tests around what’s rendered at different times. This is really nice. We’re even testing routing old URLs and making sure the correct responses are received so when traffic arrives, it’s sent to the new location.