WordPress, Next.js, or Astro? Here's How I Actually Decide for Client Projects.

"So what should we build this on?"

I get this question at the start of every single project. And every time, I resist the urge to just say "Astro" and move on. Because the honest answer is: it depends. Not in the annoying, non-committal way consultants say it. It genuinely depends on things most comparison posts never even mention.

Not which framework has the nicest docs. Not which one got the most GitHub stars this month. The real question is: who's going to maintain this thing after I hand it over? What's the actual budget, not just for building it, but for keeping it alive? And is the client ever going to touch the code themselves?

After 100+ projects across all three platforms, I've developed a framework for deciding. Here's the honest version.

The Question Every Client Forces Me to Answer

Most clients don't come to me saying "I want a Next.js app" or "Build me an Astro site." They come with a problem. "I need a website for my business." "My current site is slow." "I want to start a blog."

The technology decision is mine to make. And that's a responsibility I take seriously, because the client is going to live with that choice for years. Pick wrong and they're stuck with a site they can't update, or hosting bills that eat into their margins, or a codebase that nobody else can maintain if I get hit by a bus.

So before I even think about frameworks, I ask four questions. That's it. Four questions that tell me everything I need to know.

My Decision Framework (The Honest Version)

Here are the four factors that actually matter. Not features. Not benchmarks. The stuff that determines whether a project succeeds or becomes a maintenance nightmare.

1. Client technical ability. Can they update content themselves? Do they have a developer on staff? Or are they going to call me every time they need to change a phone number? This single question eliminates options faster than anything else.

2. Budget (build AND ongoing). Everyone asks how much a website costs to build. Almost nobody asks what it costs to run. Hosting, security updates, plugin licenses, developer retainers. The build cost is the tip of the iceberg.

3. Content update frequency. Are they publishing blog posts weekly? Updating product pages daily? Or is this a "set it and forget it" marketing site that changes twice a year?

4. Performance requirements. Is page speed critical for their business? Are they in a competitive SEO niche where Core Web Vitals matter? Or is this an internal tool where nobody cares about Lighthouse scores?

These four factors matter more than "which framework is newest." Let me walk you through how each option stacks up.

WordPress: Still the Default for a Reason

I know, I know. I literally wrote a post about why I stopped using WordPress after 6 years. But here's the thing: I stopped using it for my own projects. I still recommend it for certain clients. Because WordPress isn't about what I want. It's about what works for the client.

WordPress powers 42.5% of all websites in 2026. That number has plateaued, and it actually dipped for the first time this year. But 42.5% means something. It means every freelancer on the planet knows it. It means there's a plugin for everything. It means your client can Google any problem and find an answer.

The real strength of WordPress is client independence. A non-technical business owner can log in, change text, upload images, publish blog posts, all without calling their developer. Try that with a Next.js app.

But let's be honest about the costs. A properly maintained WordPress site runs $5,000 to $15,000 per year when you factor in hosting, premium plugins, security monitoring, and a developer retainer for updates. Plugin conflicts are real. Security patching is constant. And the "just install a plugin" mentality leads to bloated sites with 23 plugins, half of which haven't been updated in two years.

Best for: Clients who need to self-manage content, tight initial budgets, e-commerce with WooCommerce, sites where the client might switch developers down the road.

Next.js: Powerful, But Are You Ready for the Bill?

Next.js is a fantastic framework. I use it for complex web applications, dashboards, anything that needs server-side rendering or heavy interactivity. I built several client platforms on it and the developer experience is genuinely great.

But there's a conversation nobody's having honestly enough: the cost.

Vercel, the company behind Next.js, offers hosting that's beautifully integrated. But their pricing model is a trap for small teams. $20 per seat per month on the Pro plan. Bandwidth overage at $0.15 per GB after 1TB. One developer's Vercel bill hit $46,485 after a traffic spike on what was essentially a static site.

I've written about my own Next.js build killing my server and how I fixed it by self-hosting with Coolify. Self-hosting is absolutely possible, and a $17.99/month VPS can handle what Vercel charges $500+/month for at scale. But self-hosting adds operational complexity. You're now responsible for deployments, SSL, monitoring, the whole stack.

Then there's the self-hosting rabbit hole. It saves money, but it costs time. For a solo agency, that's a real tradeoff.

Best for: SaaS products, complex web applications with authentication and dynamic data, dashboards, projects where you need SSR. Not for: A marketing site, a blog, a portfolio. That's like using a Formula 1 car to go grocery shopping.

Astro: The New Kid That Grew Up Fast

In January 2026, Cloudflare acquired the Astro team. The entire team. All full-time Astro employees are now Cloudflare employees. The framework stays open-source, MIT-licensed, but now it has the backing of one of the biggest infrastructure companies on the internet.

That acquisition changed everything for me. It means Astro isn't going anywhere. It means tighter integration with Cloudflare's edge network. And it means the hosting story got even better.

Astro generates static HTML by default. Zero JavaScript shipped to the browser unless you explicitly need it. The result: sites that are 2-3x faster than equivalent Next.js builds and 50-80% cheaper to host. Cloudflare Pages' free tier gives you 500 builds per month and unlimited bandwidth. Unlimited. For free.

I migrated three client sites from WordPress to Astro earlier this year. The performance gains were immediate. The hosting costs dropped to essentially zero.

The catch? Astro requires a developer. There's no drag-and-drop editor. No plugin marketplace. If the client needs to update content without calling you, you need a headless CMS behind it, which adds complexity and cost.

Best for: Marketing sites, blogs, portfolios, content-heavy sites where performance and SEO matter. Any project where the developer handles updates.

The Real Cost Breakdown (Numbers, Not Vibes)

Let's put actual numbers on this. Not "it depends" ranges. Real costs based on what I've seen across my projects.

WordPress:

• Build: $0 to $5,000 (theme + plugins + customization)
• Annual running cost: $5,000 to $15,000 (hosting, plugins, security, dev retainer)
• Three-year TCO: $15,000 to $50,000

Next.js:

• Build: $8,000 to $25,000 (custom development)
• Annual running cost: $240 to $6,000 (self-hosted VPS vs Vercel Pro)
• Three-year TCO: $8,720 to $43,000

Astro:

• Build: $5,000 to $15,000 (custom development)
• Annual running cost: $0 to $500 (Cloudflare Pages free tier or minimal hosting)
• Three-year TCO: $5,000 to $16,500

The numbers speak for themselves. For content sites, Astro's total cost of ownership is dramatically lower. For complex applications, Next.js justifies the higher hosting costs. WordPress wins when you need a client to manage content independently, because the alternative is paying a developer for every small change, which adds up fast.

84% of organizations cite cost as their biggest cloud challenge (Flexera 2025). These aren't theoretical numbers. They're the difference between a project that's sustainable and one that bleeds money.

My Actual Decision Tree

After all that context, here's the simplified version. The decision tree I actually use.

Does the client need to edit content themselves with zero developer help? WordPress. Nothing else comes close for non-technical content management.

Are you building a web app with authentication, dynamic data, and complex interactivity? Next.js. It's built for this. Don't fight it.

Is it a marketing site, blog, portfolio, or content site where performance matters? Astro. Every time.

Does the client have a big budget and want the best of both worlds? Headless WordPress as the CMS, Astro as the frontend. Content editors get the WordPress admin they know. Visitors get blazing-fast static pages. Everybody wins.

That's it. Four questions, four answers. Everything else is noise.

What I'm Actually Recommending to Clients Right Now

In 2026, Astro is my default recommendation for most new client projects. Content sites, marketing pages, blogs, portfolios. The Cloudflare acquisition sealed it. The performance is unbeatable, the hosting is essentially free, and the developer experience is excellent.

WordPress for clients who absolutely need full self-service content management. I still set up WordPress sites, and I probably will for years. There's nothing wrong with it when it's the right tool.

Next.js only when we're genuinely building something that needs it. A dashboard. A SaaS product. Something with real-time data and complex state management. Not a marketing site with a contact form.

The framework wars are exhausting. Astro fans will tell you it's perfect for everything. WordPress developers will defend it to the death. Next.js evangelists act like you need server components to display a phone number.

The truth is boring: they're all good tools. The question isn't which is "best." It's which is best for this specific client, this specific budget, and this specific project.

That's the answer I give every time someone asks. It's not exciting. It doesn't go viral on Twitter. But it's honest. And honest is what I'd want if someone was making a technology decision that I'd have to live with for the next three years.

Need help figuring out the right stack for your project? Let's talk. I'll give you the same honest framework, no upsell attached.