Headless commerce is often presented as the obvious modern choice. It is a real architecture with real benefits and a real bill, and plenty of businesses would be better served by a hosted storefront.
What headless actually buys you
Front-end freedom. You control markup, routing and rendering completely. Custom product configurators, unusual merchandising, an editorial storefront that does not look like a store — all straightforward, versus fighting a theme system.
Performance ceiling. A well-built Next.js storefront with static product pages, edge caching and properly sized images will beat most themed platforms. Ceiling is the operative word: it is achievable, not automatic. A careless headless build is slower than a good hosted theme.
Composability. Commerce engine, CMS, search and reviews each chosen on merit and swapped independently. Valuable when one piece is strategically important — say a search experience that drives your conversion rate.
Omnichannel. One commerce API serving web, mobile app, kiosk and marketplace. If you genuinely need several channels, this is the strongest argument on the list.
What it costs you
You now own the front end. Everything a platform gave you free — cart, checkout flows, account pages, address validation, search, filtering, SEO output, accessibility — is yours to build and maintain. This is consistently underestimated.
Checkout is the hard part. Most teams keep the platform's hosted checkout precisely because PCI scope, fraud tooling, wallets, local payment methods and tax are genuinely difficult. That means a redirect out of your beautiful custom storefront, which is worth deciding on deliberately rather than discovering late.
Integrations no longer just work. The app ecosystem that installs in a click on a hosted platform assumes the platform's front end. Headless, each one is an integration project.
Operational surface. Hosting, caching, build pipelines, preview environments, monitoring, on-call. You need someone who owns it.
Content editing regresses unless you invest in it. Merchandisers lose visual editing and get a CMS plus a deploy. Solvable — visual editing, previews — but it is work, and the complaints start immediately.
When a hosted storefront wins
Choose hosted if most of these hold:
Your catalogue and merchandising are conventional
You have no dedicated front-end team
Time to market matters more than differentiation
You depend on platform apps
Revenue does not yet justify ongoing front-end engineering
A well-configured hosted theme, a good CDN and disciplined image handling get you a fast, working store in weeks. Many businesses going headless would convert better by spending the same budget on product photography, site search and page speed.
Architecture options
If headless is right, the common shapes:
Commerce API + Next.js. Shopify (Storefront API / Hydrogen), commercetools, Saleor or Medusa behind a Next.js front end. Catalogue and orders stay in the commerce engine; you own presentation. Checkout usually hosted.
Commerce API + CMS + Next.js. Add Sanity, Contentful or a custom CMS for editorial. The integration question is which system owns the product page — usually commerce owns price and inventory, CMS owns story and layout. Decide explicitly, or you get two sources of truth for the same field.
Fully custom. Your own catalogue, cart and order model. Full control, and you have chosen to build tax, fraud and payment reconciliation yourself. Justified for unusual models — licensing, subscriptions with odd proration, B2B pricing tiers — rarely otherwise.
Whichever you pick, get the caching right: product pages static or revalidated, price and inventory fresh (or fetched client-side on the detail page), cart always dynamic. Serving a stale price is a customer-service problem, and an out-of-stock add-to-cart is worse.
Migration path
Do not rebuild everything at once.
Instrument the current store first. You need a real performance and conversion baseline, or you cannot tell whether the rebuild helped.
Move the least risky surface first — content pages, landing pages, the blog. Real traffic, no revenue risk.
Then category and product pages, where performance gains are largest. Keep the existing cart and checkout initially.
Cart and checkout last, if at all. Many teams stop here permanently, and that is a legitimate end state.
Run both in parallel behind a routing rule, and watch conversion by template rather than in aggregate.
Migrate URLs carefully: redirect every old product and category URL to its new one, keep structured data, and watch Search Console for coverage errors weekly. Botched URL migrations lose more revenue than any performance gain recovers.
The honest test: can you name a specific thing your storefront must do that the hosted platform cannot? If yes, headless is probably right. If the answer is "it would be faster and more modern", you are describing an outcome achievable more cheaply.