E-commerce Search and User Guidance: Concepts, Techniques, and Product Design

The article begins by emphasizing that a product's search function is a critical pathway for users to quickly access desired information, directly influencing user experience and commercial revenue in large e‑commerce platforms.

It defines user guidance as helping users reach their goals efficiently, and in the e‑commerce context, it means assisting users to find and purchase the right products.

Core e‑commerce search objectives are listed: clarifying user intent, saving search time, improving experience, supporting a healthy ecosystem, and driving higher revenue.

The user search process is broken into three stages—search before, search during, and search after—each with specific product features and technical solutions.

Search Before

Focuses on entry placement (top tab, central search bar, sticky header) and pre‑search recommendations such as default suggestions, hot queries, and personalized bottom‑layer prompts. Technical approaches range from simple manual configuration to statistical models (hot terms) and more complex NLP pipelines (POS tagging, NER) that extract keywords and categories from recent user behavior.

Search During

Describes real‑time query autocomplete, suggestion, and correction. Simple prefix matching, Trie‑based top‑K retrieval, and statistical frequency models are discussed, as well as advanced models using embeddings (seq2seq, LSTM) and retrieval‑based ranking that combine user, context, and query vectors.

Query rewriting techniques (query‑to‑query) are presented to bridge the gap between user input and product descriptions, employing embedding similarity, multi‑method fusion (behavioral, session, content), and machine‑learning models (LR, GBDT).

Search After

Addresses result presentation, including intelligent correction, filtering, handling of no‑result scenarios, and ranking. It outlines static scoring (product quality, conversion) and dynamic scoring (query relevance, personalization) using learn‑to‑rank models, with features such as click‑through rates, user demographics, and context.

The article also compares implementations across major Chinese platforms (Taobao, JD, Pinduoduo, Tmall), highlighting differences in UI placement, suggestion styles, and ranking strategies.

In conclusion, the piece stresses that search technology must evolve together with product design, leveraging data analysis to iteratively improve each stage of the user journey.