Shopify themes or headless? Five questions that settle it

The headless conversation has calcified into two slogans. One says themes are a straitjacket and headless is how serious brands build. The other says headless is an expensive way to rebuild what Shopify already gave you. Both are repeated with more confidence than either deserves, usually by people who haven’t looked at the store in question.

The useful version of this decision isn’t a verdict. It’s an honest accounting of what each option hands you, what it quietly hands you responsibility for, and the handful of questions that tend to settle it either way.

What headless actually gives you

Headless separates the storefront from the commerce backend. Shopify keeps handling products, orders, payments, fulfilment and — importantly — checkout. The customer-facing site becomes yours.

Control that a theme genuinely can’t reach. Your own rendering strategy, your own component model, content from wherever you like, and a front end that isn’t negotiating with someone else’s section schema. If your buying journey involves configurators, quote flows, gated trade pricing or a catalogue that doesn’t behave like a catalogue, this stops being aesthetic preference and becomes the requirement.

One codebase across many storefronts. This is where headless earns out fastest. Several regions or several brands, served from shared components with local overrides, beats maintaining forked themes that drift apart within a year.

The tooling has matured. Hydrogen is no longer the experiment it was at launch. It now sits on React Router 7 with Vite, on a quarterly calendar-versioned release cadence, with Oxygen as a first-party edge runtime included with your Shopify plan rather than a separate hosting bill. Ten years of “roll your own frontend and good luck” has become a supported path.

Content and commerce in one system. If you have an editorial team that expects a real CMS and a merchandising team that expects real products, headless lets both have their proper tool instead of forcing one into the other’s.

What Shopify themes give you, which is more than it was

The strongest argument against headless isn’t that it’s hard. It’s that the thing it’s replacing keeps getting better.

Themes are a different product now. Shopify’s Horizon theme arrived in May 2025 and is the default for new stores, and the architecture underneath it changed: theme blocks, web components, and far deeper metaobject and metafield support. A lot of the “the theme won’t let me” complaints we heard three years ago describe a Shopify that no longer exists.

You inherit performance rather than owning it. With a theme, the platform’s rendering and CDN are doing work you’d otherwise be accountable for. Go headless and Core Web Vitals become yours in full — including INP, which replaced First Input Delay in March 2024 and wants interactions answered within 200ms at the 75th percentile. A framework does not hand you that. Careful engineering does.

The app ecosystem assumes a theme. This is the single most underestimated cost. A meaningful share of Shopify’s value is apps that install by dropping into a theme. Headless means each one either offers a headless SDK, exposes an API you’ll integrate yourself, or stops being available. Auditing that list against your actual installed apps is the most useful hour you can spend before committing, and it’s the step most often skipped.

Someone edits the site on a Tuesday afternoon. If that’s a merchandiser working in the theme editor today, headless means building them a replacement before launch, with preview. That’s not a footnote — it’s a workstream, and forgetting it is how a fast new storefront ends up with a slower content process than the one it replaced.

Headless doesn’t make a site fast. It makes the site’s speed your responsibility. Those are different sentences, and only one of them is a feature.

The Shopify checkout deadline sitting over both options

Worth separating from the architecture question, because it applies either way: Shopify’s migration off checkout.liquid is still running. Plus stores lost the ability to edit legacy Thank You and Order Status pages on 28 August 2025, and non-Plus stores have until 26 August 2026 to complete the same upgrade to Extensions in Checkout.

At the time of writing that’s a month out, and there’s no grace period on the other side of it: the legacy pages stop being editable and customer data stops reaching them. If you’re weighing a replatform, the checkout work isn’t optional and doesn’t get cheaper by being folded into a bigger project. Do it on its own timeline.

The five questions

Is the storefront actually the constraint? Often it isn’t. The real problem is search returning nothing useful, images nobody has compressed, or eleven apps executing on every page. All three are fixable inside a theme for a fraction of a replatform. We’ve run that engagement several times and the client kept their theme and got most of the win.

How many storefronts in two years? One store in one market is headless’s weakest case. Four brands across three regions is its strongest.

Can the theme express the journey? Not “would it be nicer bespoke” — can it express it at all? Configurators, gated pricing, B2B quoting and unusual bundling are where the answer becomes a straight no.

Who owns it in year two? A headless storefront is a software project with the maintenance profile of one. With an in-house developer or a retained partner, fine. Without either, you’ve swapped a platform problem for an abandonment problem.

What’s the appetite for surface area? Redirects, canonicals, structured data, sitemaps per collection and variant, preview environments, the build pipeline and the reason a deploy failed at 4pm on a Thursday. All perfectly manageable. All yours.

One thing we don’t treat as a trade-off: keep Shopify’s checkout either way. It’s the highest-stakes surface on the site and the one with the least to gain from being custom.

So: themes or headless?

A conventional catalogue in one market with a small team is usually best served by an excellent theme build and the boring fixes — and that’s a real answer, not a consolation prize. Multiple regions, a genuinely unusual buying journey, or an in-house team the theme is now slowing down: that’s when headless pays for itself rather than just costing more.

Most stores we look at sit clearly on one side. The ones that don’t are the interesting conversations.

---

We build both, which is the only reason this piece could stay neutral. Our Shopify practice covers custom theme builds and headless storefronts on Hydrogen, plus apps and extensions, migrations, integrations, and the performance and CRO work that frequently turns out to be what a store actually needed. An agency that only builds one of those has a strong incentive to discover that yours needs it.

If you’re weighing this — or you’re on a non-Plus plan that still hasn’t moved its Thank you and Order status pages — tell us about the store. You’ll get the reasoning above applied to your catalogue, your installed apps and your team, and we’re happy for the answer to be six weeks of fixes rather than a rebuild.