WordPress in 2025: Why It Still Powers 43% of the Web

WordPress in 2025: Why I Still Build Everything on It

WordPress powers 43% of the web and I’m responsible for, like, 0.0001% of that. But it’s my 0.0001% and I will defend it with an irrational loyalty that probably says more about me than the platform.

Here’s the thing: WordPress taught me how to be a developer. Not a WordPress developer - a developer. And after 16 years of building on it, freelancing on it, managing servers for it, and occasionally cursing at it, I still think it’s the best choice for most of what I build. Let me tell you why, and then let me be honest about the parts that make me want to throw my laptop into the Hudson River.

How WordPress Taught Me Everything

I didn’t go to college. I dropped out of high school, got my GED, and spent 11 years working in hospitality. The spark for my entry into development was an HTML4 for Dummies book I found on a free bookshelf at a coffee shop. I read a single line about wrapping “hello world” in a heading tag and opening it in a browser, did exactly that, and it was like a hit of drugs. From there I taught myself from Google. I built my first site - this site, 211j.com - in GoDaddy’s tiny modal editor. No line numbers. No syntax highlighting. Just me and raw HTML in a box the size of a sticky note.

When I found WordPress, I didn’t just install it and start making pages. I read the source code. Line by line. I needed to understand how it worked, the same way some people need to take apart a clock to understand time. And what I found inside WordPress was an architecture that shaped how I think about code to this day.

The hook system - actions and filters - is the single most important programming concept I’ve ever learned. The idea that you can define a point in your code where other code can run, without the original code knowing or caring what gets hooked in? That’s beautiful. That’s extensibility without coupling. That’s how WordPress lets thousands of plugins coexist without (usually) blowing each other up.

I took that pattern everywhere. I built an Electron boilerplate that uses action and filter hooks. My custom WordPress themes use hooks instead of hard-coded logic. My entire Claude Code expert system - the AI-powered development environment I’m building right now - is architecturally inspired by WordPress hooks. The pattern is just that good.

Why I Choose It Over Frameworks

I work at an agency as a senior front-end developer. We build Electron apps. I spent nine months on a shell application. At work, it’s React all day - that’s the house framework. I know React. I can write it. I prefer Vue, personally, but that’s a different rant.

For my freelance work? WordPress, every time. And the reason isn’t sentiment. It’s pragmatism.

WordPress has the largest market share of any CMS by a massive margin. Know what that means in the age of AI? Every AI model is ridiculously well-trained on WordPress code. When I use Claude Code to build WordPress themes, plugins, or custom functionality, the output quality is excellent because the training data is deep. The AI has seen every pattern, every hook, every template tag, every edge case. Try getting the same quality of AI assistance for a niche framework with 2% market share. It’s not the same.

Beyond AI, the ecosystem is unmatched. Need e-commerce? WooCommerce. Need forms? Gravity Forms. Need caching? LiteSpeed Cache, W3 Total Cache, pick your poison. Need custom fields? ACF. Every common problem has a battle-tested solution. You’re not reinventing wheels unless you want to.

And the client handoff story is better than anything else I’ve used. My freelance clients can log into WordPress and update their own content without calling me. They can write blog posts, swap images, update business hours. Try handing a Next.js site to a small business owner and telling them to “just push to the repo.” That’s not a handoff, that’s a hostage situation.

Managing My Own Stack

I’ve run my own server for the better part of 16 years. Linode VPS, cPanel essentially the whole time, with LiteSpeed added more recently. I kept my hosting roster deliberately small - a handful of clients, every one of those sites WordPress. I manage the stack top to bottom: server configuration, PHP versions, MySQL optimization, SSL, backups, security hardening, the whole deal.

People tell me I should use managed WordPress hosting. Kinsta, WP Engine, Flywheel. And honestly? They’re fine. For most people, they’re great. But I’m not most people. I’m the guy who reverse-engineers things because he can’t use something he doesn’t understand. I want to know why the site is slow, not just that it is. I want to configure the server for my specific workload, not accept whatever defaults the managed host decided were good enough.

Running my own server also means my hosting business - uphost.ly - has margins that make sense. I’m not paying $30/site/month to a managed host and reselling at $40. I’m paying for a VPS and running a small, sustainable roster on top of it. The economics work because I keep it intentionally small and I have the skills to manage the infrastructure. WordPress made those skills worth developing.

The Honest Problems

WordPress has real problems and I’m not going to pretend it doesn’t.

The editor situation. Gutenberg is… a thing that exists. The block editor has come a long way, but it’s still awkward for developers who want precise control over markup. Full Site Editing is ambitious and half-baked. The classic editor was simple and predictable. Gutenberg is powerful and unpredictable. I’ve made my peace with it, but “made my peace” is not the same as “enthusiastic.”

Security by popularity. When you power 43% of the web, you’re the biggest target. WordPress core is actually well-maintained security-wise. The problem is plugins. Every plugin is a potential vulnerability, maintained by someone who may or may not be keeping up with security best practices. I’ve cleaned hacked sites. It’s never fun. It’s almost always a plugin.

PHP’s reputation. People love dunking on PHP. I love PHP. I think modern PHP (8.x) is a legitimately good language that gets a bad rap because of what PHP 4 looked like. But the stigma is real, and it affects hiring and perception. When a React developer hears “WordPress” they think “template tags and spaghetti code.” They’re wrong, but perception matters.

Performance takes work. A default WordPress install is not fast. Making it fast requires caching, image optimization, CDN configuration, database optimization, and a server stack that isn’t shared hosting running Apache on a 2015 Intel Xeon. WordPress can be fast. But it doesn’t come that way out of the box.

Why It’ll Be Here in 2030

WordPress isn’t going anywhere. Not because it’s the best technology - it’s not - but because it has something no other platform has: an ecosystem that’s too deep to replace. Millions of developers, thousands of plugins, agencies built entirely on it, hosting companies built entirely for it. That’s not a product. That’s an economy.

And honestly? The AI angle is going to make WordPress even more dominant. As AI tools get better at writing code, the platform with the most training data wins. WordPress is the training data. Every Stack Overflow answer, every tutorial, every plugin source code, every theme - it’s all in the models. Building on WordPress in 2025 means you have the most capable AI assistants available for your specific stack. That’s not a small advantage.

I started with an HTML book from a free shelf. WordPress gave me a career, an architecture philosophy, a freelance business, and a hosting company. It’s messy, opinionated, sometimes infuriating, and absolutely the right tool for what I do. If you’re building for the web and you haven’t seriously considered WordPress, you’re leaving the best-supported, most flexible, most AI-friendly platform on the table.

And if you have considered it and decided it’s not for you - fair enough. But I’ll be over here with my hooks and my filters and my 43% market share, building things that actually ship.