How Yanxuan Designs a Multi‑Layer Inventory System for Seamless E‑Commerce

Basic Methods of E‑Commerce Inventory Management

E‑commerce has evolved from simple inventory‑procurement‑sales systems to comprehensive ERP, inventory management, and order management solutions.

Core Problems Solved by Inventory Management

Modern omni‑channel, multi‑warehouse strategies require inventory systems to:

Support real‑time warehouse operations (inbound, outbound, stock‑taking).

Provide decision data for order fulfillment across warehouses.

Flexibly accommodate complex sales models (single/multi‑channel, regional, activity‑based, online/offline).

E‑Commerce Inventory Layered Architecture

Warehouse Layer

Manages raw stock data in WMS or store systems, tracking SKU, location, batch, expiration, etc.

Physical Layer (Real‑stock)

Aggregates warehouse data for order scheduling. It records SKU quantities at warehouse, region, or total levels without detailed location info.

Example: SKU X has 2 units in Warehouse A and 2 in Warehouse B. A purchase of 1 unit reduces the sellable stock to 3, while the physical stock remains 4 until the order is allocated.

Sales Layer

Exposes inventory to customers. Derived from the physical layer but may differ in granularity (region + SKU, SKU only, activity‑specific). It drives user‑facing availability.

Yanxuan Inventory Center Design Practice

Business Characteristics

Yanxuan operates multiple online channels (Yanxuan APP, enterprise purchase, Taobao, Tmall, JD) and offline stores, sharing inventory across them. Dozens of promotional activities run concurrently, each with distinct stock needs. Warehouses are distributed nationwide, and any warehouse can ship to any region, eliminating regional stock partitions. Warehouse operations are outsourced to third‑party WMS providers, so Yanxuan does not directly control warehouse‑layer data.

Layered Design of Yanxuan Inventory Center

Because direct warehouse data is unavailable, Yanxuan focuses on two layers:

Sales Layer – simple SKU‑level stock for order fulfillment.

Physical Layer – enriched with batch, expiration, and location dimensions to support internal processes such as expiration tracking and batch‑level allocation.

The two layers are linked by inventory formulas so that changes in physical stock automatically propagate to sales stock.

Lock‑Stock Design

Beyond promotions, many scenarios need reserved stock (e.g., third‑party platform rules, channel‑specific pre‑stock, B2B contract reservations, offline store replenishment). Yanxuan introduces a “locked stock pool” concept.

All shared inventory forms a large pool; within it, independent locked stock pools act as subsets.

Each locked pool has a unique lockkey . Users request a locked pool by providing lockkey + SKU + quantity. Once allocated, they consume stock using only the lockkey, enabling isolated management for activities, contracts, or special channels.

This design simplifies handling multiple concurrent promotions and other special‑use cases, but currently suffers from weak pool governance, inconsistent lockkey standards, and higher integration complexity.

Conclusion

Effective e‑commerce inventory management relies on a layered data model (warehouse, physical, sales) and a supply‑chain strategy that minimizes cost and stockouts while ensuring rapid fulfillment. Yanxuan’s approach demonstrates how a focused layered design and flexible lock‑stock mechanisms can support complex, fast‑growing retail operations.