SAP Commerce for the Chinese Market: A Different Project, Not a Different Locale

Most market rollouts on SAP Commerce are configuration: a language, a currency, a price catalog, a payment method, done. China is the market where that mental model fails completely. The world's largest e-commerce market runs on different platforms, different payments, different search engines, different analytics, different logistics expectations, and a legal regime that dictates where your servers physically sit. The old CX Works series on China said it plainly and it remains true: entering China is a project of its own, with its own architecture. This guide consolidates that series and updates the legal and technical landscape.

Market Structure: The Duopoly Shapes Everything#

Tmall (Alibaba) and JD dominate Chinese B2C, with Alibaba's properties also controlling C2C (Taobao) and a large share of B2B (1688). Two structural consequences for your strategy:

• Customer expectations are set by the giants. UI conventions (utility icons on the right edge, QR-code login from the app instead of passwords), the cart model, delivery speed, and after-sales generosity are all calibrated to Tmall and JD. Your storefront is judged against them, not against your global site.
• Marketplace-first is a legitimate architecture. For brands building presence, a Tmall or JD store frequently precedes (or permanently replaces) an owned platform, because the marketplaces bring the traffic. The realistic SAP Commerce role in that phase is order and inventory backbone behind marketplace channels via integration, not storefront. Owned-platform investment makes sense once brand traffic justifies it, and the two models commonly run in parallel.

Peak events are structural too: Singles' Day (November 11) and 618 (June) dwarf Western peaks in intensity. The high-traffic guide's entire discipline applies, plus two China-specific practices from the source material worth keeping: dedicated event inventory linked to event promotions through your sourcing strategy (so event orders ship fast and do not cannibalize regular stock), and business-plus-IT standby staffing through the event window with content ready for live adjustment.

Law First: Residency, Filing, Moderation#

Nothing here is legal advice; everything here has ended projects that ignored it.

• Data residency. The Cybersecurity Law's data-localization principle, extended since the source era by the Data Security Law (2021) and the Personal Information Protection Law (PIPL, 2021), means data generated serving the Chinese market is expected to stay in China, with regulated pathways for any cross-border transfer. Architecturally this is the instance-strategy decision from the regions guide: a China deployment is typically a separate instance in a Chinese data center, not a site on your global estate. Verify current SAP Commerce Cloud availability for China with SAP during planning; historically the answer has involved dedicated arrangements or self-managed deployments on China-region cloud infrastructure, and the answer changes over time, so get it in writing for your timeline.
• ICP filing. Websites served to mainland China must complete ICP registration before going public, and commercial operation adds licensing requirements tied to a local legal entity. This gates your domain, your CDN (a China-local CDN requires the ICP filing), and your launch date; start it early because the clock is bureaucratic, not technical.
• Consumer law. Seven-day no-reason returns for most online purchase categories are a legal baseline, and advertising them prominently is also a trust signal customers expect. Wire the returns flow and the policy copy together.
• Content moderation. User-generated content (reviews, comments, images) must be screened before public display; politically sensitive or illegal content on your platform is your liability. Budget a moderation pipeline (typically a third-party moderation API plus human escalation) in the reviews feature, not as an afterthought.
• Blocked services. Google, Facebook, YouTube, and most Western analytics, fonts, CDNs, and tag managers do not function in the mainland. Every third-party script on your storefront needs a China audit; one blocked synchronous script can hang page loads for the entire market.

Hong Kong and Taiwan are different regimes entirely (Traditional Chinese, different currencies, no Great Firewall constraints); model them as separate sites in your multi-site setup, not as variants of the mainland site.

UX Differences That Reach the Data Model#

Two source-era observations that remain sharp because they touch commerce mechanics, not styling:

• The cart is a wishlist with checkboxes. Chinese customers park items in the cart for weeks awaiting promotions and select a subset at checkout. That means selectable cart entries (a selected flag your checkout honors), retargeting logic that does not treat a full cart as imminent purchase, and cart-abandonment campaigns recalibrated accordingly.
• Order placement and payment are separate steps. The customer submits the order, then pays (Alipay/WeChat redirect or QR scan), and unpaid orders expire after a window. Your order state machine needs a first-class PENDING_PAYMENT state with expiry-and-release logic for inventory, which is not how card-authorization checkouts are usually modeled.

Add QR-code login (scan with the app to authenticate the web session), aggressive mobile-first assumptions, and WeChat as a first-class channel, and the storefront requirements read like a new head in the headless guide's sense, which is exactly the right frame.

Channels: WeChat Is Not Optional#

The channel inventory beyond web and app: WeChat Official Accounts (menu-driven brand presence), WeChat mini programs (app-like experiences inside WeChat, now a standard commerce channel), self-service QR ordering in physical contexts, and mobile POS. Technically the good news is architectural: a mini program is an OCC consumer like any other head. The composable storefront serves your web presence; the mini program consumes the same APIs with WeChat's auth woven in. Build the commerce APIs once, let each channel be a head.

Payment: Alipay, WeChat Pay, UnionPay#

Card-centric checkout assumptions do not survive contact with China. Alipay and WeChat Pay dominate, with UnionPay covering bank-integrated cases. Integration realities:

• All the major gateways offer full APIs including refund automation; the refund-via-API flow matters because the seven-day return right makes refunds a volume operation.
• Payment aggregators that wrap Alipay, WeChat Pay, and UnionPay behind one integration are the pragmatic default, trading a fee margin for one integration instead of three and faster onboarding.
• Model the QR-and-redirect flows properly: payment confirmation arrives asynchronously (notify callbacks), which pairs with the PENDING_PAYMENT order state above. Idempotent callback handling and reconciliation jobs are the difference between a payment integration and a finance incident.

Social login follows the same logic: WeChat, QQ, and Weibo cover effectively the entire online population, and login-with-WeChat dramatically outperforms registration forms.

Search, Analytics, and App Distribution#

• SEO means Baidu. The structural differences from the source still apply: slower, less transparent indexing of young domains (site-level authority weighs more than page-level), weaker original-source detection (syndication can outrank you), and a preference for frequently updated sites. Solid technical SEO carries over; timeline expectations do not. Plan months, not weeks, for organic visibility.
• Analytics means Baidu Tongji (web and app SDKs, including mini-program support), because Google Analytics does not load. The analytics guide's event-model discipline transfers; the tooling does not.
• App distribution: Apple's China App Store for iOS (with regional catalog differences), and a fragmented Android landscape of manufacturer stores (Huawei, Xiaomi, Oppo, Vivo) plus third-party stores in place of Google Play. Pick the top handful by your audience; nobody ships to all of them.

Logistics: The Two-Day Bar#

JD's same-day/next-day logistics set the national expectation; the practical bar for a credible independent platform is roughly two-day average delivery. The design consequences from the source hold up: position warehouses east and split north/south (the demand mass is eastern), integrate third-party logistics rather than building delivery (SF Express and the Cainiao network carry most of the market), and expose live tracking in the storefront because customers check it obsessively. Click-and-collect, beloved in Europe, is marginal here; fast home delivery beat it.

Customer care runs through chat, deeply integrated: agents expect the customer's profile, order history, and return calculations in one view (ASM plus your service tooling), and chat is simultaneously the pre-sales advisor, the after-sales desk, and your feedback pipeline.

The Architecture Summary#

A China entry on SAP Commerce typically means: a separate instance in-country (residency), an ICP-filed domain on a local CDN, a storefront rebuilt against local UX expectations (or a mini-program-first strategy), a payment aggregator integration with async confirmation flows, Baidu for search and analytics, 3PL integration with tracking, a content moderation pipeline, and marketplace integrations run as peer channels. Roughly none of your global estate's edge configuration survives the trip, most of your platform code does, and the honest planning stance is the one this guide opened with: a different project, scoped and staffed as one, sharing your competence but not your assumptions.