Why we stopped building WordPress sites for most clients

But I always hated the architecture underneath. The plugin ecosystem in particular, a messy sea of dependencies, each one a potential conflict, a performance hit, a security risk. We spent a lot of time stripping plugins back out of client sites we'd inherited. The better we got at WordPress, the more we were compensating for its weaknesses rather than building on its strengths.

The move to a component-based stack was a natural progression from where we already were. If you're building flexible content blocks in ACF, you're already thinking in components you're just doing it in a much less elegant environment. When we started working with Sanity and Next.js, the logic carried straight over. The difference was that everything underneath it was cleaner, faster, and more purposeful.

Performance was my main reason for making the jump. Not in an abstract sense, in the sense that a WordPress site with a reasonable plugin load is fighting against itself from the start, and there's a ceiling on what you can achieve without significant ongoing maintenance. The sites we build now are faster by default, not by effort.

But the bigger reason was where we wanted to take things. We wanted to build AI into our sites that had real business context, not generic chatbots, but tools that understand your content, your services, your customers. We've already done that with our analytics tools built into Sanity — and WordPress recently released MCP capability, I think they'll get there eventually, but the architecture isn't built for it. We wanted to build it in now, not bolt it on later.

The thing that kept us on WordPress longer than we probably should have stayed was client anxiety. "What happens if you go away? Who'll manage the site?" It's a fair question and we still hear it when we're showing our new tech stack to our clients. Our answer: you own everything. Your content, your codebase, your data. If you wanted to take it to another developer tomorrow, you could.

That used to feel like a harder sell than it does now. More developers are working across different tech stacks, partly because AI tooling makes it easier to get up to speed quickly. The barrier to "what if I need someone else" is lower than it's ever been. We think that makes the switch easier to justify than it was two years ago.

We still maintain a handful of WordPress sites for existing clients and we're not going anywhere on that. But for new builds, we're not going back. The gap between what we can offer on our current stack and what we could offer on WordPress is too wide.