Supply-Demand Modeling and Category Optimization for the Idle Second-Hand Market

Background: Rising traffic costs make fine‑grained operation essential; in the idle second‑hand market, individual sellers' unique items require category optimization to improve efficiency and drive growth.

Product Understanding: Listings are often unstructured, so extracting and structuring key attributes is the foundation for category optimization.

Category Decision Tree: User decision factors are mined from search, interaction, and transaction data, matched to product attributes, and weighted by conversion rates to rank key attributes.

Key Attribute Structuring: Because many attributes are missing, the product side guides users to complete structured information; NLP, image recognition, and similar‑item algorithms supplement missing data, while the data team monitors coverage and quality.

Supply‑Demand Mining & Opportunity Market: Using the decision tree and structured attributes, a supply‑demand index is built from four sub‑models (price, inventory, search‑hotspot, demand‑activation) to assess market status.

Price Model: Prices reflect supply‑demand balance; seasonal examples show how price trends indicate excess or shortage.

Inventory Model: Compares segment efficiency with overall market using transaction volume, traffic, and item counts, defining a nine‑grid of market states (high/medium/low efficiency) to diagnose opportunities.

Search‑Hotspot Model: Identifies high‑search‑volume items with low supply, e.g., a specific mask brand.

Demand‑Activation Model: Evaluates the time from listing to transaction to uncover fast‑moving supply.

Application: The index guides category adjustments—boosting supply for scarce categories and using search and price guidance for over‑supplied ones—to drive business growth.

Future: Ongoing optimization of product understanding and index application will be shared in upcoming technical articles.