How to Turn Complex E‑commerce Promotion Logic into Drag‑and‑Drop Blocks

Death Case: A Comma Caused Million‑Level Loss

A cosmetics platform’s promotion rule "满300-50，适用品类：护肤，彩妆（不含口红）" missed a closing parenthesis, causing the system to interpret it as "护肤彩妆且不含口红". High‑price items like serums could not receive the discount, resulting in a 2 million‑yuan GMV loss and a 300% increase in complaints.

Root Causes : The rule engine required operators to write raw code expressions, offering no real‑time preview and only generic "system error" messages.

Six‑Step Democratized Design

1. Visual Rule Engine: From Code to Blocks

Goal : Enable operators to configure complex logic via drag‑and‑drop without understanding Boolean operators.

Componentization Atomic components: product category, price range, user tags. Logical relations: AND, OR, NOT. Action modules: discount, gift, points.

Logic Visualization Canvas supports drag‑linking to generate a logic tree similar to a flowchart. Real‑time rendering with color‑coded hierarchy (e.g., red for conflicts).

Live Simulator Enter test order amount/product to auto‑calculate discount results. When errors occur, the system highlights the offending condition node (e.g., "price > 100").

Example: Configuring "满300‑50, 叠加新人券, 排除特价品" becomes four drag‑and‑drop steps: order amount ≥300, user tag = new, exclude product tag = special price, then set priority.

2. Natural Language Engine: Let the System Speak Human

Goal : Allow operators to write conditions in plain language, automatically translating them into machine‑readable rules.

Semantic Parser Pre‑defined templates such as "商品属于【类目】且【价格】≥X元". Fuzzy input correction (e.g., "不包括口红" → "排除商品类目=口红").

Bidirectional Translation Natural language ↔ technical rule ↔ SQL, synchronized in real time. Operators can switch views to verify logic accuracy.