I’ve been mulling this idea over for a while and I’m going for it. Over the next few weeks I will rebuilding RobertJacobi.com with as much core Gutenberg and as few plugins as technically possible. With the update to WordPress 5.7 happening March 9, 2021, it’s as good a time as any to pursue this experiment.
This is sweet and simple, load WordPress sites in as fast as 37ms with Cloudways!
Even in the most tightly managed WordPress sites, some cruft is bound to get into the mix. I have 19 active plugins and honestly that seems like way more than there should be. Some are flat out redundant, a couple of optimization plugins that certainly could be handled by one good plugin. So with that I should be able to go from 3 to 1. Now we are at 17 plugins. There are two plugins for statistics, one which could simply be embedded in a theme, and the other being JetPack.
Which actually brings us to Jetpack as whole, this is a powerful suite of services but my host, Convesio, covers some of the basics with automated backups and anti-spam protection via the included HumanPresence functionality. Jetpack stats are pretty good but that’s already covered. Jetpack does add a number of very useful blocks but unless I need to expand the functionality, I think I can at least, in the short term, run without Jetpack. Now we’re down to 16 active plugins.
There are three random “utility” plugins that can definitely go, two for some UI/UX which is no longer being used – yes I’ll have to find the posts where they used to be used but there are very few and I hope that core blocks can solve those issues from last year. Another is an SSL redirect which I honestly don’t even know why is still active, I’m pretty sure this is already managed at the host.
The last utility plugin cleanup will be removing WP Page Permalink Extension. This hasn’t been updated in a year, and it’s time for me to get over the novelty of keeping .html at the end of pages and posts. That will of course mean making sure that I don’t blow up any SEO juice I’ve built up and managing redirects correctly, but I have faith in the universe that those issues will be minor. Look at that, now we are at 13 active plugins.
I do have two plugins from the GoDaddy Pro universe which will get removed at the initial relaunch but I expect at least one of them to come back very quickly. These two are ManageWP and CoBlocks. With a focus on core, CoBlocks should become redundant so I’ll have to clean up any content using those blocks. ManageWP is connected to The Hub by GoDaddy Pro. I think there is a lot of interesting stuff going on at The Hub, and I’ll want to take advantage of the journey. At least in the very short term, I’ll be down to 11 active plugins.
What am I looking to add? In the short term I have only one plugin on the radar. Restrict Content Pro. I plan on having some gated content and services and quite simply, this looks to solve the problem with minimal effort.
At the moment I’m using Newspack by Automattic, available on GitHub. I like the clean look for a news and opinion based site, it’s not official so autoupdates don’t occur, and I’ve been really itching to try Oxygen. I’m pushing for something as simple as possible as a starter but with support, flexibility, and performance as a goal. Now will this potentially add blocks and plugins that aren’t core? I don’t know yet. The claim is “zero bloat” so it will be an interesting experiment.
What about building from scratch? I’ve gone this route, and surprisingly the performance I expected wasn’t there. So much happens on the host, CDN, etc. that there is a fine line between a really good theme, and the most insane and tight code possible. There have been a number of good talks recently about the environmental impact of running a website so I want to be as cognizant of anything superfluous in a theme but there’s also the impact of me spending much more time on developing, testing, and debugging – and any bugs that I could accidentally inject that leave me open to security issues.
The End Goal
The ultimate über end goal is complete performance (the speed of the site AND the speed of utilization). I want all code rot completely eliminated. I want less to manage. I want a site where I can evaluate new plugins and solutions on a very stable base. I like testing on a real site not just a stock install, so this is important. Also it will be interesting to measure against GTMetrix and Supervisor the improvements made.
Looking forward to any comments and thoughts – there will definitely be a post mortem as completion of this project.