The client wants a "headless e-shop". After an hour-long conversation we usually find out they want either SaaS with better speed, or a fully custom platform. Headless is somewhere in between — and sometimes that's the right place, sometimes it isn't.
This article is about when it makes sense to go "headless" and when it's an over-engineered solution for needs that Shopify would handle.
What "headless" actually means
An e-shop has two layers: 1. **Backend / commerce platform** — product catalog, cart, checkout, orders, customers. Traditionally: Shopify, WooCommerce, Magento. 2. **Frontend / store** — what the visitor sees. Traditionally: a built-in theme from the platform (Shopify Liquid, WooCommerce theme).
Headless = separation of these layers. The backend stays (or is swapped for an API-first platform like Medusa, Commerce.js, Saleor), the frontend is built as a separate application (Next.js, Astro, SvelteKit) that communicates with the backend via API.
When headless makes sense
1. The market demands speed that a built-in theme can't deliver.
Lighthouse Performance score above 90, LCP under 2 s, FID under 100 ms on mobile. Shopify Liquid with 30 apps + 10 pixel trackers will never get close to this. Next.js with ISR and edge-rendered catalog will.
If your business case is "2% conversion → 3% conversion through performance", headless serves the purpose. If your business case is "95% conversion is enough", it doesn't.
2. The frontend needs its own logic that isn't primarily e-commerce.
Example: B2B e-shop with a 3D product configurator (custom furniture). A Shopify theme won't deliver a 3D viewer. Custom React/Next.js will.
Second example: AI-driven personalization where the frontend changes the offer based on real-time behaviour. Headless enables this, a built-in theme doesn't.
3. Multi-channel — web, mobile app, kiosk — from one backend.
If you want one catalog, one checkout, but two or three different frontends, headless is the only reasonable solution. Maintaining two e-shops in sync is a horror.
4. The team has React/TypeScript engineers who own the frontend.
Headless = custom code. Without an engineering team (or external partner), the system won't grow. Shopify Liquid can be tuned by a senior marketer. A Next.js storefront can't.
When headless doesn't make sense
1. The market is "standard e-shop, 50–500 products, B2C".
Shopify with a Dawn or Impulse theme will sell more for you than custom Next.js. The theme has features (reviews, related products, search, checkout) ready. Custom = you have to build every feature.
If your requirements match 95% of e-shops, custom is over-engineered.
2. Budget is smaller than €30,000.
Headless minimum: backend (Medusa.js hosted = €200–500/month), frontend (Next.js development = €15,000–40,000 for the first phase), CMS for content (Sanity, Strapi = €200–500/month), pipeline (Vercel/Cloudflare = €100–300/month), maintenance (€2,000–5,000 monthly).
If the business doesn't have €30k+ upfront and €2k+ monthly operations, Shopify Starter or Basic is the right choice.
3. The team isn't technical and never will be.
Headless = constant technical decision-making. Without an internal team / external partner guiding you, the system stalls on "how to resolve a change in checkout".
How to decide
Three questions:
**Question 1:** What is your GMV (gross merchandise volume) for the last 12 months? - < €100k → Shopify Basic, don't think about headless - €100k–500k → Shopify with a headless theme, or non-headless Shopify Plus - €500k–5M → Headless starts to make sense - > €5M → Headless is often a necessity
**Question 2:** Do you have 1+ engineer with React/TypeScript on the team? - Yes → headless is a realistic option - No → either find an external partner, or stay on a built-in platform
**Question 3:** Do you plan B2B, a mobile app, or kiosk in 2 years? - Yes → headless is an investment that returns at expansion - No → stay on a simpler platform
What stack we use
For the Slovak / Central European market in 2026 we recommend: - **Backend**: Medusa.js (open-source, self-hosted, fully controlled) - **Frontend**: Next.js 15+ on Vercel/Cloudflare - **CMS**: Sanity.io (content + product enrichment) - **Search**: Algolia (cost → Meilisearch self-hosted) - **Analytics**: Plausible or PostHog (GDPR-friendly) - **Payments**: Stripe + GoPay (Slovak market) - **Email**: Klaviyo or Postmark
This stack serves €1M–20M GMV without a rewrite. With further growth, parts get replaced gradually (Medusa → custom microservices, Algolia → in-house search).
---
*We help clients choose architecture before they commit to a technology. We often recommend Shopify if headless doesn't make sense — it depends on the business case, not on technological preference.*