Headless commerce is no longer "a Vercel thing". In 2024 there are at least three serious ways to ship an e-commerce with a custom frontend. Here they are.
Shopify Hydrogen
Hydrogen is Shopify's official Remix-based framework for custom storefronts. Hosting on Oxygen (Shopify's edge platform) or Vercel. Pros: mature Shopify backend — inventory, payments, tax, shipping all solved. Cons: total lock-in to Shopify pricing; Oxygen has some limits for global traffic.
Headless WooCommerce
WooCommerce behind WPGraphQL with a Next.js frontend. Pros: no lock-in, predictable cost, full control. Cons: WooCommerce was not built headless and it shows — checkout especially is a job if you want to replace it. For most Italian clients with mid catalogues (1-5k SKUs) it works great if you accept keeping standard checkout.
Medusa
Open source, Node.js, modular. Conceptually the cleanest: headless-by-design backend, Next.js frontend. Pros: no licences, deploy anywhere (Hetzner self-hosted included), REST and GraphQL APIs. Cons: young ecosystem (plugins less mature vs Woo), needs a Node-fluent team.
Our matrix
| Case | Pick |
|---|---|
| Consumer brand, strong design, high budget | Shopify + Hydrogen |
| Italian SME with ERP already integrated | Headless WooCommerce |
| Marketplace or B2B custom with unique flows | Medusa |
| Simple e-commerce, limited time/resources | Monolithic WooCommerce |
What does not matter
"Performance" is not the main reason to go headless in 2024. A well-cached monolithic WooCommerce (LiteSpeed, FastCGI cache) runs fine. You go headless when brand and custom UX justify the complexity — not because you want a 5-point higher Lighthouse score.