Headless e-commerce has been the most over-marketed concept in retail tech for about six years now. Every platform sells it. Every agency pitches it. Most brands who buy it would have been better off on a normal Shopify theme. A smaller group of brands genuinely need it and benefit enormously.

Telling the two groups apart is the whole job. This post is the honest version — what headless actually is in 2026, what an Australian implementation realistically costs, and the specific conditions under which it pays off.

My opinion up front: most brands do not need headless. The ones that do need it know exactly why, and they are usually trying to do something a templated platform structurally cannot do.

What headless means in plain language

A "headed" store ships the storefront and the commerce engine as one product. Shopify is the canonical example. When you edit a Liquid theme, you are editing the head of the system — the part the customer sees. The cart, the checkout, the products, the orders all live on the same platform, accessed through the same admin.

Headless cuts the head off. The commerce engine — products, cart, checkout, inventory, orders — stays in one system. The storefront becomes a separate application, built in React or Next.js, that talks to the commerce engine through an API.

The two pieces communicate but they are no longer welded together. You can swap the head without touching the body. You can swap the body without touching the head. You can build any frontend experience you want without being constrained by the platform's theming system.

That is the entire architectural claim. Everything else is a consequence.

The three flavours of headless

Calling something "headless" without specifying which flavour is mostly meaningless. There are three meaningfully different setups.

Headless Shopify

You keep Shopify as the backend. Products, cart, customers, checkout — all still in the Shopify admin. You build a custom frontend in Shopify Hydrogen or Next.js that fetches data through the Storefront API. Cart and checkout happen on Shopify's infrastructure.

This is the most common flavour. It is the lowest-risk path to a custom frontend. You keep the admin your team already knows. You keep Shopify Payments and Shop Pay. You lose nothing operationally on the back end.

Composable commerce on an open-source backend

You replace Shopify entirely with an open-source commerce engine — usually Saleor (GraphQL, Python/Django, very mature) or Medusa (Node.js, modular, growing fast). The storefront is a custom Next.js app. Checkout is custom-built on Stripe.

This is the highest-effort path and the highest-control path. You own the entire stack. You can build any commerce flow you want. You pay no platform fee. You take on more engineering responsibility.

Composable on a SaaS backend that is not Shopify

A small middle ground — using commercetools, BigCommerce, or Swell as the headless backend with a custom frontend. Less common in the Australian market because the local agency and developer pool skews heavily Shopify or Saleor.

The case for headless

I want to make the case before the criticism.

When you have a content-led brand and a templated frontend is killing your conversion

If your brand thesis is design and storytelling — premium fashion, design objects, food, anything where the visual brand carries the buy — and you are stuck inside Shopify's section model, headless gives you what you actually need. You can build any layout you want, any animation, any product experience. The frontend stops being the limit.

When you genuinely need a multi-channel API

If you sell on web, in a native app, on in-store POS terminals, on social commerce, and on B2B portals — and they all need to share the same products and inventory — headless is the model that supports that. The commerce backend becomes a service every channel talks to.

When performance is a real conversion lever

A well-built headless Next.js storefront on Vercel will outperform any Liquid theme on the same hosting tier. Largest Contentful Paint under 1.5 seconds, Interaction to Next Paint under 200ms, full edge caching. Plus has caught up considerably with Hydrogen and Oxygen, but a hand-tuned headless frontend still wins on the metrics.

Studies consistently show a one-second improvement in load time correlates with conversion lifts of 5% to 7%. For a brand doing $10M AUD/year, a 5% conversion lift is $500,000 in incremental revenue. The build pays for itself.

When app-stack sprawl has become the actual problem

If your Plus stack has 22 apps and three of them duplicate functionality, the headless rebuild is usually the cleanup. You write the features you actually need into the codebase. You delete the apps. You stop paying $4,000/month in subscription overhead.

The honest case against headless

For balance — the reasons most brands should not do it.

It is more expensive to build and more expensive to run

A real headless build for an Australian brand sits between $80,000 and $200,000 AUD one-off, depending on scope. Maintenance is $5,000 to $15,000 AUD/month for ongoing engineering. Hosting is $1,500 to $4,000 AUD/month if you do it properly.

For a brand doing under $3M AUD revenue, the math basically never works.

You lose the app store

Shopify's app store is one of its real strengths. Need a new feature? Install an app for $49/month. On headless you build the feature, or you find a way to integrate the SaaS version of it. Either is slower than installing a Shopify app.

Your team's velocity depends on engineering

On Plus your marketing team can ship a new collection landing page in an afternoon. On headless, they file a ticket. A good headless build mitigates this with a CMS layer that gives marketers control over content blocks — but the day-to-day velocity is genuinely different.

Checkout is a problem worth respecting

Shopify spent over a decade tuning their checkout. It works. Replacing it with a custom Stripe checkout is doable and the Stripe domestic Australian rate of 1.75% + A$0.30 (1.925% + A$0.33 after GST) is meaningfully cheaper than Shopify Payments on Plus — but you are taking on the responsibility of optimising checkout conversion yourself. If you do it badly, you will lose more in conversion than you save in fees.

Recovery from failure is your problem now

On Shopify, when the platform has an outage, every Shopify store has the outage. It is collective infrastructure failure that you read about on TechCrunch. On headless, when your custom backend has an outage, only your store has the outage. The on-call rotation, the postmortems, the monitoring tooling — these become your team's responsibility. Most brands underestimate the operational maturity required to run their own commerce stack reliably.

What it really costs an AU brand in 2026

Let me put numbers next to the architecture.

Headless Shopify (custom Next.js frontend on Shopify backend)

Build: $50,000 to $120,000 AUD
Frontend hosting (Vercel): $200 to $800/month
Continue paying Shopify Plus
Engineering retainer: $4,000 to $10,000/month

You keep almost all the operational simplicity and gain a custom frontend. Reasonable starter for brands at $3M to $8M AUD revenue who want a brand-led storefront without changing their backend.

Full composable (Saleor + Next.js + Stripe)

Build: $120,000 to $250,000 AUD
Hosting and infra (AWS Sydney, Vercel): $2,000 to $5,000/month
Engineering: 1 in-house or strong retainer, $10,000 to $18,000/month
Stripe processing: 0.4% to 0.6% cheaper than Plus
No platform fee

Real annual operating cost: $200,000 to $300,000 AUD for the stack. Same as a $4M+ brand pays Shopify Plus all-in. Suitable for brands at $8M+ where the build pays back in 18 to 24 months and the strategic control matters.

The Australian context that matters

A few facts to ground the decision.

The Australian Privacy Act reforms are removing the $3M turnover small-business exemption in late 2026, which means every Australian e-commerce brand will be covered by the 13 Australian Privacy Principles. Custom builds make the disclosure and consent obligations easier to comply with — you control exactly what data is collected and how. SaaS stacks with 20 apps means 20 data-processor relationships to document.

Stripe Australia's domestic card rate is the second-lowest in the world after the UK and EU, thanks to Reserve Bank of Australia interchange regulation. The processing math for custom checkouts in Australia is better than in almost any other market.

AusPost MyPost Business shipping integration is straightforward through their Parcel Send API — no platform is required. Your custom build talks to AusPost directly. So does Aramex, Sendle, and the rest.

The local engineering market for headless commerce talent is thin but real — Sydney, Melbourne and remote-AU developers with Saleor and Hydrogen experience are findable but not abundant. Plan accordingly.

When an AU brand should genuinely care

You should be researching headless if any of these are true:

You are doing $5M+ AUD revenue and growing
Your brand thesis is design-led and Shopify's section system is genuinely limiting you
Your app stack is above 15 apps and three of them are doing things you would prefer to own
Performance is a material conversion lever for you, not a vanity metric
You have or are willing to hire one mid-level engineer for ongoing maintenance

If three of those five apply, run the spreadsheet. The build probably pays back.

If none of them apply, stay on Plus. Spend the money on inventory and ads.

The honest bottom line

Headless is a serious architecture for serious brands. It is not a checkbox upgrade. It is not a fashion trend. It is a deliberate decision to take more control of your storefront in exchange for taking on more responsibility.

For the brands who need it, it is one of the highest-impact investments they can make. For the brands who do not need it, it is a six-figure mistake that gets blamed on the agency.

A useful starting point: drop your storefront URL into our free audit. The report covers the rendered mobile and desktop performance, the SEO surface, and the specific issues that are most likely sending revenue away. If your current Plus setup is already doing well on those numbers, the headless rebuild won't add 30%. If it's bleeding, you'll know which problem you're actually trying to solve.

Headless E-commerce: Real 2026 Cost for AU Brands