Every few weeks someone in Yorkshire emails me a version of the same question. "We've been told we need headless WordPress. What is it, and do we?" Usually an agency has pitched it, the quote was big, and they want a second opinion from someone who doesn't gain either way. Here's what I tell them.
What headless WordPress actually means
Normal WordPress does two jobs. It's where you write and store your content, and it's the thing that builds the page a visitor sees, using a theme. Headless splits those. WordPress keeps the first job. A separate front end, usually built in something like Nuxt or Next, takes over the second, pulling your content out of WordPress through its API and rendering the pages itself.
You still log in to the same admin. You still write posts in the same editor. The difference is that the theme is gone, and the public site is a separate application, often hosted somewhere else entirely.
If the word headless is new to you, the plain English version of the wider idea is on my other site. That covers headless as a category. This post is about the WordPress flavour specifically, because that's what people in Hull and Leeds keep getting quoted for.
Why anyone bothers
Speed is the usual reason. A well built headless front end serves pre-rendered pages from a CDN, so there's no PHP running and no database query on each visit. That can be very fast. I say can, because I've seen headless builds that were slower than the WordPress site they replaced, and a hand-built theme on a proper performance budget gets most sites into the high nineties on Google's speed tests without any of this.
Security is the second. If the public site is static files, there's nothing for the usual WordPress bots to hammer on the front. WordPress itself can sit behind a login on a separate address that nobody but you knows about.
The third is the one I think is strongest: you need WordPress content in more than one place. A website and a mobile app. A site and a screen in reception. A PWA that has to work with no signal. One source of content, several front ends. That's where I've done most of my headless WordPress work, and it's the case where it earns its cost.
The honest downsides
Two systems to look after instead of one, and two hosting bills. When something breaks you have to work out which side it's on before you can fix it.
Plugins that touch the front end stop working. Forms, SEO plugins that write tags into the page, page builders, galleries, most of them assume WordPress is rendering the page. You rebuild that behaviour in the front end or find a different route. This catches people out more than anything else on the list.
Preview gets harder. In normal WordPress you click preview and see the page. Headless, somebody has to wire that up, and a lot of the headless builds I've inherited never did, so editors were publishing blind.
And it costs more to build. Not double, but noticeably more, because the front end is a real piece of software rather than a theme. I'd rather you knew that before the quote lands than after.
When it's worth it
Your site is slow and you've already done the normal fixes. Most slow WordPress sites don't need headless. They need fewer plugins, smaller images and decent hosting. I'd sort that first, and it costs a fraction of a rebuild.
You need the same content in a website and an app, or on a device with a poor connection. This is the strongest case and it's usually a clear yes.
Traffic spikes hurt you. A static front end shrugs off a spike that would flatten a shared host, which matters if you get featured somewhere or run a big campaign.
You've got a developer who'll still be around in two years. Headless with nobody to maintain it is worse than plain WordPress, because there are two things to go stale instead of one.
When it isn't
A twelve page site for a business in Beverley with a contact form and a news page doesn't need headless. A hand-built theme on a performance budget will be fast, the owner can edit it themselves, and there's one bill and one login. That's most of the WordPress work I do, and it's the right answer more often than agencies would like.
If what you're really after is an online shop, headless WordPress is probably the wrong tree. Headless WooCommerce is a serious project with its own set of problems, and there are better routes for a store. That side of things lives with Headless Digital, my ecommerce brand, so I'll point you there rather than pretend I'd take it on here.
How I build it when it is the right call
Nuxt on the front, because it's what I've used every day for years and it handles static generation, server rendering and PWA features without a fight. WordPress stays on a normal host, locked down, feeding the front end through the REST API or WPGraphQL. The front end deploys to Vercel or similar, rebuilds when content changes, and I wire up preview on day one so editors can see a draft before it goes live.
A lot of the repetitive parts, the content fetching, the image handling, the build and deploy setup, come from tooling I built myself. That's why I can do a headless WordPress build in weeks rather than the months you may have been quoted, and why the running cost afterwards is low. Efficient, not cut-price. The output is what I'd ask you to judge.
A Yorkshire note
I'm in Hull. Most of my clients are in Hull and East Yorkshire, with a few up towards York and Leeds. Nothing about headless needs a local developer. But when there are two systems and a question of which side a problem is on, being able to sit in the same room for an hour and draw it on a whiteboard is worth more than it sounds. And when I say headless isn't worth it for your site, you know I'm not saying it from a sales office two hundred miles away.
Get a second opinion before you sign
If you've been quoted for headless WordPress and you're not sure, book a free 30 minute surgery. I'll tell you if it's the right call for your site, and if it isn't, what would actually fix the problem you started with. Either way you'll leave knowing what you're buying. If it is the right call, this is how I do WordPress and we can talk about what it would take.
Free 30-min surgery
Working on something similar?