필사 모드: E-commerce Platforms 2026 — Shopify (Hydrogen) / Medusa / Saleor / Vendure / WooCommerce / BigCommerce Deep Dive
English> "The most expensive thing in e-commerce is not the platform license — it is the cost of migrating off the wrong platform four years later." — paraphrased Shopify Unite 2025 keynote note, 2025
Picking an e-commerce platform in 2026 is no longer a simple SaaS-signup decision. It is an architectural decision that shapes the next five years of the business. Even Shopify alone has fanned out into Hydrogen (a React Router 7 storefront), Oxygen (edge runtime), Polaris (admin design system), and the Hydrogen Visual Editor (Shopify 2025 Editions). On the other side, Medusa just closed a Series A, Saleor matured in Poland, and Vendure built a TypeScript stronghold.
As of May 2026 the platform space splits into five buckets — **SaaS / headless (open source) / all-in-one / drop-in / regional** — and each bucket has further forks ("Hydrogen vs. Liquid?", "Medusa Cloud vs. self-host?"). This article puts 22 platforms on one page and ends with a "who should pick what" matrix.
1. The 2026 e-commerce platform map — SaaS / headless / open source / regional
Here is the entire platform space in one table.
| Category | Representative products | Core value |
|---|---|---|
| Integrated SaaS | Shopify, BigCommerce, Adobe Commerce (Magento) | Admin + payments + hosting + marketplace in one |
| Headless SaaS | Shopify Hydrogen, BigCommerce Catalyst, commercetools | They provide the API, you build the front |
| Open-source headless | Medusa, Saleor, Vendure, Crystallize, Swell | Self-hostable, you own the core code |
| All-in-one (no code) | Wix, Squarespace, Webflow + Wized | Mouse-built stores |
| WordPress ecosystem | WooCommerce (Automattic), WP EasyCart | Blends with blog/CMS |
| Drop-in checkout | Snipcart, Foxy.io, Lemon Squeezy | Bolt checkout onto an existing site |
| Korea-local | Cafe24, Makeshop, Naver SmartStore, Toss Seller, Coupang Wing | Local PG, tax, and logistics integration |
| Japan-local | BASE, STORES, Shopify Japan, MakeShop Japan, KARTE Commerce | Friendly with JP payments (konbini, PayPay) |
The classification axis is one line: "where does the platform stop being responsible, and where do you start writing code?" Wix and Squarespace own almost everything but leave you little room to customize. Medusa and Vendure put the core in your hands, but every uptime page is also yours.
**Three biggest shifts in 2026**: (1) Shopify now ships Liquid, Hydrogen, and the Visual Editor as a single bundle, so the "Hydrogen or Liquid?" binary is gone. (2) After its Series A, Medusa stabilized v2.0's modular architecture and Medusa Cloud, breaking the old "open source = ops burden" formula. (3) Vercel pushed the NextCommerce template plus the AI Commerce Toolkit, cementing the Next.js camp's default starting point.
2. Shopify + Hydrogen 2025 + Oxygen + Polaris — the dominant leader
Shopify is still the clear #1 in 2026. Per Shopify's own filings, 2M+ stores worldwide process more than 292B USD in annual GMV. The real strength is not the single SaaS — it is the **layered stack**.
- **Storefront layer**: Liquid (classic themes), Hydrogen (React Router 7), Hydrogen Visual Editor (2025 Editions)
- **Edge runtime**: Oxygen (Shopify's managed Workers-like edge)
- **Admin / design system**: Polaris (React component library, v13 in 2024)
- **APIs**: Storefront API (GraphQL), Admin API (REST + GraphQL), Bulk Operations API
- **Payments**: Shopify Payments (Stripe-powered), Shop Pay (one-click)
- **Extensibility**: Shopify Functions (Rust/WASM), App Bridge, Flow, Hydrogen Component Library
The headline Hydrogen 2025 change is the **React Router 7 migration**. Up through 2024 Hydrogen built on Remix v2, but Remix was folded back into React Router 7 and Hydrogen moved with it. A fresh Hydrogen project now looks like this.
// app/routes/products.$handle.tsx — Hydrogen 2025 + React Router 7
export async function loader({ context, params }: { context: any; params: any }) {
const { storefront } = context
const { product } = await storefront.query(PRODUCT_QUERY, {
variables: {
handle: params.handle,
selectedOptions: getSelectedProductOptions(new Request('')),
},
cache: storefront.CacheLong(),
})
if (!product?.id) throw new Response(null, { status: 404 })
return { product }
}
export default function Product() {
const { product } = useLoaderData<typeof loader>()
return (
)
}
const PRODUCT_QUERY = `#graphql
query Product($handle: String!, $selectedOptions: [SelectedOptionInput!]!) {
product(handle: $handle) {
id
title
description
}
}
`
Oxygen is, in practice, a Cloudflare-Workers-like edge that Shopify operates. Shopify Plus customers get Oxygen for free; lower tiers self-host on Vercel, Netlify, or Cloudflare. Oxygen's selling point is that **GraphQL Storefront API latency drops to roughly 1–2ms**, so an average Hydrogen page lands a 1.2–1.8s LCP comfortably.
Pricing starts at Basic 39 USD/month, Shopify 105 USD/month, Advanced 399 USD/month, with Plus from 2,300 USD/month. Shopify Payments charges roughly 2.4–2.9% per transaction; using an external PG adds another 0.5–2%. There are 45,000+ Plus customers globally as of 2026.
3. Medusa — open source headless (Series A)
Medusa launched in 2019 out of Copenhagen, Denmark. It raised an 8M USD seed in 2022 and a roughly 11M USD Series A in 2025, becoming the flagship of the open-source camp. The GitHub project has crossed 28k stars in 2026.
The core story is the **v2.0 modular architecture**. The single-monorepo days are gone; Medusa is now a set of independent modules.
- `@medusajs/medusa` — core backend (Node.js + TypeScript)
- `@medusajs/admin` — admin UI (Vite + React)
- `@medusajs/cart`, `@medusajs/order`, `@medusajs/customer`, and other domain modules
- `@medusajs/types` — shared type definitions
- Starter: `@medusajs/medusa-starter-default` (Next.js 14)
The differentiator is the **Region model**. Multi-language, multi-currency, multi-tax, multi-shipping all live in one backend through Region as a first-class object. For example a Japan Region might be JPY + 10% consumption tax + Japan Post; a Korea Region could be KRW + 10% VAT + CJ Logistics.
// medusa-config.ts (Medusa v2)
export default defineConfig({
projectConfig: {
databaseUrl: process.env.DATABASE_URL,
http: {
storeCors: process.env.STORE_CORS!,
adminCors: process.env.ADMIN_CORS!,
authCors: process.env.AUTH_CORS!,
jwtSecret: process.env.JWT_SECRET!,
cookieSecret: process.env.COOKIE_SECRET!,
},
},
modules: [
{ resolve: '@medusajs/medusa/cache-redis', options: { redisUrl: process.env.REDIS_URL } },
{ resolve: '@medusajs/medusa/event-bus-redis', options: { redisUrl: process.env.REDIS_URL } },
{ resolve: '@medusajs/medusa/workflow-engine-redis', options: { redis: { url: process.env.REDIS_URL } } },
{ resolve: '@medusajs/medusa/payment-stripe', options: { apiKey: process.env.STRIPE_API_KEY } },
],
})
Medusa Cloud removes the ops burden of self-hosting and is priced at Starter 25 USD/month, Pro 99 USD/month, with Enterprise on request. Self-hosted, Medusa boots on PostgreSQL + Redis + one container; roughly 50 USD/month of infrastructure handles 100k orders/month.
The biggest weakness is **no theme/design system**. The one-click theme swap of Shopify is impossible — every UI element ships as part of your own Next.js storefront. The official Next.js starter is solid enough to ship a PoC inside a week, however.
4. Saleor — Polish open source
Saleor is an open-source headless e-commerce platform built by Saleor Commerce, a Polish company. The stack is Python + Django + GraphQL. As of 2026 the GitHub project has 21k stars and the company has taken roughly 80M USD in funding. The defining trait is **GraphQL-first**: the entire API is GraphQL, with REST limited to a few webhooks.
Saleor GraphQL example — product lookup
query ProductDetails($slug: String!) {
product(slug: $slug) {
id
name
description
pricing {
priceRange {
start {
gross {
amount
currency
}
}
}
}
variants {
id
name
quantityAvailable
}
}
}
Saleor ships as:
- **Saleor Core** — Python/Django GraphQL server
- **Saleor Dashboard** — React admin (Next.js 14)
- **Saleor Apps** — micro-service style extensions (Stripe, SendGrid, etc.)
- **Saleor Storefront** — Next.js 15 reference store
Enterprise customers include Lush, Breitling, Ikon Pass, and Eyewa. Managed hosting via Saleor Cloud is priced at Starter 90 USD/month, Pro 690 USD/month, and Enterprise on request. The 2025 release of the Saleor App SDK completed the marketplace story, letting third parties publish apps into the Saleor app marketplace.
Strengths: (1) GraphQL consistency, (2) multi-channel and multi-warehouse baked into the core, (3) Polish/EU origins make GDPR a non-issue. Weaknesses: (1) Python is a higher entry barrier for Node-only teams, (2) Korean and Japanese payment integrations are thin — KG Inicis, Toss, PayPay, etc. all need custom adapters.
5. Vendure — TypeScript open source
Vendure is a 100% TypeScript headless e-commerce framework that started in the UK. The stack is NestJS + TypeORM + GraphQL. As of 2026 the GitHub project has 6k stars, and while the team is small, the core is well respected for being rock-solid. Versus Medusa and Saleor, the Vendure pitch is **B2B scenarios and multi-vendor marketplaces**.
Key concepts:
- **Channels** — run multiple stores (B2C, B2B, marketplace seller) from a single core
- **Sellers** — first-class objects for multi-vendor marketplaces
- **Promotions** — rule-based promotion engine (conditions plus actions)
- **Stock Locations** — multi-warehouse inventory isolation
- **Plugins** — NestJS-module-style extensions
A Vendure plugin looks like this.
// my-loyalty-plugin.ts
@VendurePlugin({
imports: [PluginCommonModule],
providers: [LoyaltyService],
shopApiExtensions: {
schema: `
extend type Customer {
loyaltyPoints: Int!
}
`,
resolvers: [LoyaltyResolver],
},
})
export class LoyaltyPlugin {}
There is no first-party SaaS managed hosting yet. Vendure Cloud is in beta, and the official recommendation is to use partners like Northflank, Railway, or Fly.io. Self-hosted, PostgreSQL + Redis + a Node container is enough; infrastructure runs roughly 30–50 USD/month.
The standout strength is **NestJS familiarity**. If your team has used NestJS, the Vendure core code reads like your own and you can ship a domain-modifying PR with confidence. Weaknesses: (1) the ecosystem is smaller than Medusa or Saleor, (2) there are no Korean PG adapters, (3) the admin UI looks dated — though a refresh has been in flight since late 2025.
6. WooCommerce — the WordPress giant
WooCommerce is an Automattic-owned WordPress e-commerce plugin. By BuiltWith estimates, roughly 28% of e-commerce sites worldwide run on WooCommerce in 2026 — somewhere between 3.8M and 4M active sites in absolute terms. That single-platform share exceeds Shopify's 2M+ stores, so statistically WooCommerce is still the largest e-commerce platform.
Strengths: (1) the WordPress ecosystem — blog, CMS, SEO, media library all integrated, (2) tens of thousands of plugins, including almost every Korean and Japanese PG adapter, (3) **100% site ownership** — self-hosted means no data lock-in, (4) the core is free.
Weaknesses: (1) PHP + MySQL is a heavy stack, (2) security patches are the site owner's responsibility, (3) you tune caches, CDN, and the database yourself once traffic grows, (4) the WooCommerce Store API is still a bit thin for headless usage.
From 2024 WooCommerce shipped block-based Checkout and Cart Blocks as the defaults, deepening the Gutenberg integration. In 2025, Automattic launched WooCommerce Cloud, a managed hosting plan starting at 70 USD/month.
// WooCommerce + WordPress — price-change hook
add_filter('woocommerce_product_get_price', function ($price, $product) {
if (current_user_can('vip_customer')) {
return (float)$price * 0.9; // VIP gets 10% off
}
return $price;
}, 10, 2);
// Send a webhook to an external ERP when an order completes
add_action('woocommerce_order_status_completed', function ($order_id) {
$order = wc_get_order($order_id);
wp_remote_post('https://erp.example.com/orders', [
'body' => json_encode($order->get_data()),
'headers' => ['Content-Type' => 'application/json'],
]);
});
Headless WooCommerce variants include **Frontity** (discontinued in 2022), **Faust.js** (WPGraphQL-based), and **Atlas Content Modeler**. The 2025 official **Atomic Blocks API** is now the recommended path.
7. BigCommerce / Magento (Adobe Commerce) — enterprise
The two enterprise SaaS giants are BigCommerce and Adobe Commerce (formerly Magento).
**BigCommerce** announced BigCommerce Catalyst in 2024 — a Next.js 15 headless storefront template — and properly entered the headless space. Catalyst leans hard on React Server Components and combines with BigCommerce's GraphQL Storefront API to become a direct rival to Shopify Hydrogen. Pricing: Standard 39 USD/month, Plus 105 USD/month, Pro 399 USD/month, Enterprise on request. Unlike Shopify, **the default transaction fee is 0%**; external PG fees still apply, of course.
**Adobe Commerce (Magento)** has been part of Adobe Experience Cloud ever since Adobe acquired Magento in 2018 for 1.68B USD. In 2026 the Adobe Commerce Cloud stack supports headless via PWA Studio plus Commerce Storefront (built on the 2024-announced Edge Delivery Services). Licensing is GMV-tiered and typically starts north of 22k–50k USD per year. The customer base is global enterprise — Coca-Cola, BMW, Nestlé.
| Item | BigCommerce | Adobe Commerce | Shopify Plus |
|---|---|---|---|
| Entry price | 39 USD/month | 22,000 USD/year (negotiated) | 2,300 USD/month |
| Transaction fee | 0% | 0% | 2.4–2.9% (Payments) |
| Headless | Catalyst (Next.js) | PWA Studio + Edge | Hydrogen (React Router 7) |
| Admin | In-house UI | Adobe Experience Cloud | Shopify Admin + Polaris |
| Multi-store | Yes (Channels) | Strong (store views) | Markets, Plus has unlimited expansion stores |
| Key customers | Skullcandy, Burrow, Solo Stove | BMW, Coca-Cola, Nestlé | Allbirds, Gymshark, Heinz |
Enterprise selection usually hinges on (1) global multi-store needs (Adobe), (2) B2B catalog and PIM strength (Adobe), (3) Salesforce/HubSpot integration depth (BigCommerce), and (4) speed of Korean/Japanese PG connection (Shopify).
8. Wix / Squarespace / Webflow + Wized — all-in-one
The no-code camp leaders are Wix, Squarespace, and the newer Webflow + Wized combo.
**Wix** has evolved into Wix Commerce + Wix Studio + Velo (its JavaScript SDK). As of 2026 it claims 22M+ active business accounts worldwide. Pricing: Business 27 USD/month, Business Elite 159 USD/month. Wix has been pushing its own AI (Wix AI Designer) that builds a site in roughly 30 seconds.
**Squarespace Commerce** is Squarespace 7.1 + Commerce + Squarespace Pixel (its first-party tracker). Pricing: Business 23 USD/month, Commerce Basic 27 USD/month, Commerce Advanced 49 USD/month. Squarespace's edge is **template design quality**, and Squarespace AI launched in 2025.
The **Webflow + Wized** combo is the newer designer-favored pattern. Design in Webflow, then attach Wized — a no-code data-binding tool — to drive dynamic catalog and checkout flows. Webflow itself offers an Ecommerce plan (Standard 29 USD/month, Plus 74 USD/month, Advanced 212 USD/month), but for serious catalogs many teams wire Webflow to an external headless source like Foxy.io or the Shopify Storefront API.
// Wized data-binding example — Shopify Storefront API
Wized.requests.execute('fetchProducts').then((response) => {
Wized.data.r.products = response.data.products.edges.map((edge) => ({
title: edge.node.title,
handle: edge.node.handle,
price: edge.node.priceRange.minVariantPrice.amount,
}))
})
Common all-in-one weaknesses: (1) data lock-in (migrating stores is painful), (2) limited custom payment and logistics integration, (3) performance ceilings above 10k visitors/day. Strengths: (1) no design-engineering split required, (2) hosting, SSL, and CDN are included, (3) non-technical staff can edit content directly.
9. NextCommerce (Vercel) / Crystallize / Swell — the newer entrants
**NextCommerce** is Vercel's open-source starter — first published in 2020, fully rebuilt in 2024 — built on Next.js 15 plus Shopify (or BigCommerce, Saleor). It lives at the vercel/commerce repo on GitHub and has crossed 12k stars. It is more of a **reference implementation** than a product, but virtually every Next.js team building headless commerce uses it as a starting point.
// next-commerce — product card component
export function ProductCard({ product }: { product: Product }) {
return (
amount={product.priceRange.maxVariantPrice.amount}
currencyCode={product.priceRange.maxVariantPrice.currencyCode}
/>
)
}
**Crystallize** is a Norway-based headless PIM + Commerce + Subscriptions platform. It is GraphQL-first, and its defining strength is **content modeling**. Pricing: Standard 29 USD/month, Pro 269 USD/month, Enterprise on request. Publishing, magazines, and subscription-oriented businesses are the primary target.
**Swell** is a headless SaaS launched in 2017 with strong subscription and marketplace stories. Swell's differentiator is treating **Subscription as a first-class object** in the data model — a great fit for B2C subscription boxes (think Misfits Market) and for B2B SaaS that bills through its own checkout. Pricing: Standard 175 USD/month, Pro 599 USD/month.
All three share (1) well-organized GraphQL/REST APIs, (2) default Stripe/Adyen payment integrations, and (3) Netlify/Vercel/Cloudflare deployment guides. None are strong on Korean/Japanese local PG integration — those adapters have to be hand-written.
10. Snipcart / Foxy.io — drop-in
Drop-in solutions exist for the moment you want to add checkout to an existing site (a blog, a static site, Webflow, Framer) without rebuilding it.
**Snipcart** is a JavaScript cart from a Canadian company that works with one line of script. Add `data-item-*` attributes to your HTML and the cart and checkout appear automatically.
<!-- Snipcart drop-in example -->
class="snipcart-add-item"
data-item-id="tshirt-blue"
data-item-price="29.00"
data-item-url="/products/tshirt-blue"
data-item-name="Blue T-shirt"
>
Add to cart $29
Pricing is a 2% GMV fee (raised in 2025) plus a 20 USD/month minimum. Duda (a website builder) acquired Snipcart in 2024 but the product still ships independently.
**Foxy.io** is a US-built headless cart + checkout, strong for Webflow, Squarespace, and static sites that need complex tax and subscription scenarios. Pricing: free trial, Standard from 75 USD/month. The 2025 Foxy v3.0 release added a Web Components-based cart, making static-site integration much easier.
The drop-in tradeoffs are stark. (1) Checkout in 5 minutes. (2) Total design freedom. (3) Zero hosting/infra burden. In exchange, (1) catalog/inventory/order management UIs are limited, (2) admin features are thin, and (3) multi-store and B2B scenarios are not a fit. Volumes of 100–500 orders per month are the sweet spot.
11. MACH architecture — Microservices / API-first / Cloud-native / Headless
**MACH** (Microservices, API-first, Cloud-native, Headless) is the enterprise e-commerce architecture principle codified by the MACH Alliance, formed in 2020 by commercetools, Algolia, Contentstack, and others. In one line: "Stop running monolithic e-commerce; compose each feature as an independent microservice."
| Principle | Meaning | Representative products |
|---|---|---|
| Microservices | Each domain (Cart, Order, Search) is an independent service | commercetools, Spryker |
| API-first | Every capability exposed as an API | Saleor (GraphQL), Medusa (REST + GraphQL) |
| Cloud-native | Friendly with Kubernetes and serverless | Vercel, AWS Lambda, Cloudflare Workers |
| Headless | Frontend and backend decoupled | Hydrogen, Catalyst, Next-commerce |
A canonical MACH stack tends to look like:
- **Commerce core**: commercetools, Saleor, Medusa
- **Search**: Algolia, Typesense, Meilisearch
- **CMS**: Contentful, Sanity, Storyblok
- **Payments**: Stripe, Adyen, Mollie
- **CDP/marketing**: Segment, Bloomreach, Klaviyo
- **DXP/storefront**: Next.js, Hydrogen, Nuxt, Astro
MACH's strength is that **each layer can be swapped independently**. If Algolia becomes too expensive, switch to Typesense; if Contentful's UX disappoints, move to Sanity. The cost is real, though: (1) operational complexity is far above a monolith, (2) integration adds spend, (3) "MACH from day one" is almost always over-engineering. The ROI usually appears once GMV crosses about 10M USD/year.
The MACH Alliance counted more than 90 members as of 2026. In Korea, commercetools powers parts of LG H&H's and Samsung's global sites.
12. Korea — Cafe24 / Makeshop / Naver SmartStore / Toss Seller
The Korean market is a different game from the global SaaS world. Korean PGs (KG Inicis, NHN KCP, Toss Payments), Korean tax (VAT, simplified taxpayer rules), Korean logistics (CJ Logistics, Logen), and Korean marketing channels (Naver Pay, Kakao Channels) all have to land as local integrations.
| Platform | Signup cost | Transaction fee | Primary target |
|---|---|---|---|
| Cafe24 | Free | 0% (PG fee separate) | Solo founders through mid-market D2C |
| Makeshop | From 16,500 KRW/month | 0% (PG fee separate) | Fashion and beauty SMBs |
| Naver SmartStore | Free | 2% (when using Naver Pay) | Anyone, dominant for solo sellers |
| Toss Seller | Free | 2.5% cards, 2% TossPay | Sellers leaning into Toss checkout |
| Coupang Wing (marketplace) | Free | 4–12% by category | Coupang marketplace sellers |
| Interpark Sellers | Free | 5–12% by category | Books, culture, ticketing |
**Cafe24** is Korea's largest hosting + e-commerce platform, with roughly 2 million cumulative global signups as of 2026. The model is to start free and pay for payments, domains, design templates, and the app market à la carte. Beyond Korea it operates in Japan, China, Vietnam, and the Philippines, and there is a separate Cafe24 Global SaaS.
**Makeshop** (a KOLON subsidiary) is a Korean e-commerce platform centered on fashion and beauty, known for design flexibility and a strong mobile admin. In 2025 it began Japan expansion in partnership with KARTE (Plaid).
**Naver SmartStore** is effectively the default for Korean solo sellers. Naver Shopping visibility is immediate, Naver Pay checkout conversion is dominant, and "browse on Naver, pay on Naver" is the natural flow. The admin is very simple, but design freedom is the lowest of the bunch.
**Toss Seller** is a 2024 launch from Toss, deeply surfaced inside the Toss app and best optimized for TossPay checkout. By May 2026 it has crossed 300,000 cumulative sellers and is quickly absorbing solo founders and small brands.
To run Shopify in Korea you have to wire **NicePay, KG Inicis, or Toss Payments** as external PGs in place of Shopify Payments, which adds another 0.5–2% on top of the transaction fee. As a result Korean D2C brands constantly weigh "Shopify + KG Inicis external PG" against "Cafe24 D2C with full design customization."
13. Japan — BASE / STORES / Shopify Japan / MakeShop Japan / KARTE Commerce
Japan's market structure looks similar to Korea's, but the payment culture is more conservative. Beyond credit cards, **konbini payment** (paying at a convenience store), **cash on delivery**, **deferred payment**, **PayPay**, and **LINE Pay** are all everyday options.
| Platform | Signup cost | Transaction fee | Notes |
|---|---|---|---|
| BASE | Free | 3% + 40 JPY (payment fee) + 3% (service) | 6M+ solo sellers |
| STORES | Free to 9,790 JPY/month | 3.6–5% | Design-friendly |
| Shopify Japan | From 33 USD/month | 3.25–3.45% (Shopify Payments) | Global brands |
| MakeShop Japan | From 11,000 JPY/month | 3.19–3.59% | Mid-market EC |
| KARTE Commerce | Negotiated (bundled with CDP) | Negotiated | CDP + Commerce integration |
| Color Me Shop | From free | From 3.4% | GMO-operated, long-standing EC |
**BASE** is Japan's #1 free e-commerce platform with 6M+ registered sellers. The "open a store in one minute" motto absorbed the solo-seller market, and BASE Pay is its own payment product. The downside is the cumulative ~6.6% transaction fee.
**STORES** is design-quality-first and popular with designers and small businesses. STORES has its own STORES Payments PG and exposes nearly every feature on the free plan, which is its strongest pitch. In 2024 the parent company unified its brand from "hey" to STORES.
**Shopify Japan** is Shopify's Japan entity, providing a Japanese admin, Japanese payments (Shopify Payments Japan runs on top of GMO-PG), and Japanese logistics (Yamato, Japan Post) integrations. PayPay checkout became a first-party integration in 2025.
**KARTE Commerce** is Plaid's commerce module that sits on top of KARTE (a CDP). The core value is "personalized commerce driven by customer behavior data." Large Japanese brands such as JINS, parts of ZOZO, and Isetan Mitsukoshi have adopted it. The common pattern is existing commerce plus KARTE analytics, not KARTE Commerce standalone.
For global D2C brands running Shopify in Japan, the biggest friction is **konbini payment** integration. Shopify Japan removed most of that pain in 2024 by officially supporting the GMO-PG adapter, though deferred payments (NP Atobarai, Paidy) still require dedicated apps.
14. Subscription commerce — Recharge / Bold / Loop
Subscription commerce reached an estimated 200B USD global market in 2026, and Shopify, BigCommerce, and WooCommerce all support subscriptions natively or through partner apps.
**Recharge** (US) is the de facto standard subscription app, used by 80%+ of Shopify Plus customers. Pricing: Standard 99 USD/month + 1% GMV, Pro 499 USD/month + 0.5%, Custom on request. Dollar Shave Club, Native, and OLIPOP power their subscriptions on Recharge.
**Bold Subscriptions** is a direct Recharge rival from Canadian Bold Commerce. Its pitch is lower pricing (from 49 USD/month) and a more flexible business-rules engine.
**Loop Subscriptions** is a newer India-born entrant that has grown fast in the Shopify App Store, particularly popular with small D2C brands. Pricing: 49 to 199 USD/month. Some Korean D2C brands (parts of Dr. G and Innisfree) have adopted it.
The hard operational problems of subscription commerce are: (1) **failed-payment recovery** — card expirations and limit issues cause a 3–7% monthly failure rate, (2) **churn management** — cumulative one-year churn lands at 50–70%, (3) **skip/pause UX** — a self-service UI letting customers skip a cycle is mandatory, (4) **tax and regional pricing** — complexity explodes once you go multi-country.
// Recharge API — pause subscription example
async function pauseSubscription(subscriptionId: string, untilDate: string) {
const res = await fetch(
`https://api.rechargeapps.com/subscriptions/${subscriptionId}/set_next_charge_date`,
{
method: 'POST',
headers: {
'X-Recharge-Access-Token': process.env.RECHARGE_TOKEN!,
'Content-Type': 'application/json',
},
body: JSON.stringify({ date: untilDate }),
}
)
if (!res.ok) throw new Error(`Failed: ${res.status}`)
return res.json()
}
The 2026 trend is **AI-driven churn prediction plus skip recommendations**. Recharge and Bold both shipped in-house ML models that flag "likely to skip the next order" customers and proactively send discount coupons.
15. Who should pick what — solo, brand, D2C, B2B, enterprise
Now that the 22 platforms have all been covered, here is a scenario-by-scenario recommendation.
| Scenario | First choice | Second choice | Notes |
|---|---|---|---|
| Korea solo seller (100–500 orders/month) | Naver SmartStore | Cafe24 or Toss Seller | Visibility first, accept the design limits |
| Korea D2C brand (1k–10k orders/month) | Cafe24 + custom design | Shopify + KG Inicis | Korean PG fit dominates |
| Japan solo seller | BASE | STORES | Can launch in a minute |
| Japan global D2C | Shopify Japan + Hydrogen | KARTE Commerce | Konbini and PayPay essential |
| Global D2C brand (English-first) | Shopify + Hydrogen + Oxygen | BigCommerce + Catalyst | Dominant leader |
| Design-first static site + checkout | Webflow + Foxy.io | Squarespace Commerce | Snipcart is also fine |
| WordPress blog + checkout | WooCommerce | Easy Digital Downloads | Unbeatable when paired with the blog |
| Headless on Next.js from day one | Medusa + Next.js Starter | Saleor + Next.js | You own 100% of the code |
| TypeScript full-stack team | Vendure | Medusa | NestJS familiarity decides it |
| Global enterprise (10M+ annual GMV) | Shopify Plus | Adobe Commerce or commercetools | Evaluate MACH |
| Subscription box | Shopify + Recharge | Swell | Failed-payment recovery is the lever |
| B2B marketplace | Vendure | Saleor or commercetools | First-class multi-vendor support |
| Game/SaaS digital goods | Lemon Squeezy or Paddle | Stripe + your own site | Merchant of Record matters |
The **golden rules** of platform selection in 2026 boil down to four.
1. **Match your current GMV.** A solo seller doing 100 orders/month should not be building a Hydrogen headless storefront. Conversely, a D2C brand doing 100k orders/month will hit a wall on Wix.
2. **Put regional payments, tax, and logistics first.** In Korea and Japan, local PG and tax integration are always the largest friction for a global SaaS.
3. **Price the migration cost up front.** Moving from Shopify to Medusa typically takes 6 to 12 months, and data-integrity breakage is the largest single risk.
4. **Headless is a team-shape decision, not a GMV decision.** You only see the ROI on headless once you have two-plus full-time React/Next.js engineers.
The 2026 e-commerce platform market has reached a stage where **no single "right answer" exists**, and that is a healthy signal. From the Korean solo seller to the global enterprise, every stage of growth has a sharp fit. The biggest mistake is asking "which platform is the best?" — the right question is "which platform best fits the next 18 months of my business?"
References
- Shopify — [shopify.com](https://www.shopify.com/), [Hydrogen docs](https://shopify.dev/docs/storefronts/headless/hydrogen), [Oxygen overview](https://shopify.dev/docs/storefronts/headless/hydrogen/oxygen), [Polaris design system](https://polaris.shopify.com/), [Shopify Functions](https://shopify.dev/docs/apps/build/functions)
- Shopify Editions — [shopify.com/editions](https://www.shopify.com/editions), [Hydrogen Visual Editor 2025](https://shopify.dev/docs/storefronts/headless/hydrogen/visual-editor)
- Medusa — [medusajs.com](https://medusajs.com/), [GitHub](https://github.com/medusajs/medusa), [Medusa v2 docs](https://docs.medusajs.com/), [Medusa Cloud](https://medusajs.com/cloud/)
- Saleor — [saleor.io](https://saleor.io/), [GitHub](https://github.com/saleor/saleor), [Saleor Cloud](https://saleor.io/cloud), [Saleor App SDK](https://docs.saleor.io/developer/extending/apps/overview)
- Vendure — [vendure.io](https://www.vendure.io/), [GitHub](https://github.com/vendure-ecommerce/vendure), [Vendure Hub plugins](https://vendure.io/hub)
- WooCommerce — [woocommerce.com](https://woocommerce.com/), [Block-based checkout](https://woocommerce.com/checkout-blocks/), [Store API](https://github.com/woocommerce/woocommerce/tree/trunk/plugins/woocommerce/src/StoreApi)
- BigCommerce — [bigcommerce.com](https://www.bigcommerce.com/), [Catalyst](https://www.catalyst.dev/), [GraphQL Storefront API](https://developer.bigcommerce.com/docs/storefront/graphql)
- Adobe Commerce / Magento — [business.adobe.com/products/magento/magento-commerce.html](https://business.adobe.com/products/magento/magento-commerce.html), [Adobe Commerce Storefront](https://developer.adobe.com/commerce/storefront/)
- Wix Commerce — [wix.com/ecommerce](https://www.wix.com/ecommerce/website), [Velo dev docs](https://www.wix.com/velo)
- Squarespace Commerce — [squarespace.com/ecommerce-website](https://www.squarespace.com/ecommerce-website)
- Webflow Ecommerce — [webflow.com/ecommerce](https://webflow.com/ecommerce), Wized — [wized.com](https://wized.com/)
- NextCommerce (Vercel) — [vercel.com/templates/next.js/nextjs-commerce](https://vercel.com/templates/next.js/nextjs-commerce), [GitHub](https://github.com/vercel/commerce)
- Crystallize — [crystallize.com](https://crystallize.com/), [GraphQL API](https://crystallize.com/learn/developer-guides)
- Swell — [swell.is](https://www.swell.is/), [Subscriptions](https://www.swell.is/subscriptions)
- Snipcart — [snipcart.com](https://snipcart.com/), [Pricing](https://snipcart.com/pricing)
- Foxy.io — [foxy.io](https://foxy.io/), [Web Components](https://foxy.dev/web-components)
- MACH Alliance — [machalliance.org](https://machalliance.org/), [Members directory](https://machalliance.org/members)
- commercetools — [commercetools.com](https://commercetools.com/)
- Cafe24 — [cafe24.com](https://www.cafe24.com/), [Cafe24 Global](https://www.cafe24corp.com/en/main/index.php)
- Makeshop — [makeshop.co.kr](https://www.makeshop.co.kr/)
- Naver SmartStore — [sell.smartstore.naver.com](https://sell.smartstore.naver.com/), [Naver Pay](https://pay.naver.com/)
- Toss Seller — [toss.im/seller](https://toss.im/seller)
- Coupang Wing — [wing.coupang.com](https://wing.coupang.com/)
- BASE — [thebase.in](https://thebase.in/), [BASE Apps](https://apps.thebase.in/)
- STORES — [stores.jp](https://stores.jp/), [STORES Payments](https://stores.jp/payments)
- Shopify Japan — [shopify.com/jp](https://www.shopify.com/jp), [GMO-PG integration](https://help.shopify.com/ja/manual/payments)
- MakeShop Japan — [makeshop.jp](https://www.makeshop.jp/)
- Color Me Shop — [shop-pro.jp](https://shop-pro.jp/)
- KARTE Commerce — [karte.io/products/commerce](https://karte.io/products/commerce/), Plaid Inc.
- Recharge — [rechargepayments.com](https://rechargepayments.com/), [API docs](https://developer.rechargepayments.com/)
- Bold Subscriptions — [boldcommerce.com/products/subscriptions](https://boldcommerce.com/products/subscriptions)
- Loop Subscriptions — [loopwork.co](https://loopwork.co/)
- Lemon Squeezy — [lemonsqueezy.com](https://www.lemonsqueezy.com/), Stripe subsidiary
- Paddle — [paddle.com](https://www.paddle.com/), Merchant of Record
- Standard Webhooks (signature standard) — [standardwebhooks.com](https://www.standardwebhooks.com/)
- BuiltWith e-commerce stats (2026) — [trends.builtwith.com/shop](https://trends.builtwith.com/shop)
현재 단락 (1/328)
Picking an e-commerce platform in 2026 is no longer a simple SaaS-signup decision. It is an architec...