Understanding Taxonomies

If you’re building e-commerce systems and encountering product classification for the first time, understanding taxonomies is essential. This guide explains what product taxonomies are, why they’re challenging to work with, and how automated classification tools like Product Classifier handle them.

Part 1: What Are Product Taxonomies?

A product taxonomy is a standardized classification system that divides the entire product space into a hierarchical tree of categories. Think of it as a massive, organized map of every possible product type.

Tree Structure

Taxonomies organize categories as nodes in a tree. Each node represents a product category, and nodes are connected in parent-child relationships. The deeper a node is in the tree, the more specific it becomes in narrowing down the product space.

Category Paths and IDs

Categories are identified by both an ID and a path that shows the route from the root to that specific node. Here’s an example from the Shopify taxonomy:

el-17 : Electronics > Video

This contains:

• The category ID (el-17)
• The category path (Electronics > Video)

The path shows the hierarchy: Electronics is the root category, and Video is a subcategory within it.

Leaf Nodes vs Non-Leaf Nodes

Take a look at these related categories from the Shopify taxonomy:

el-17 : Electronics > Video
el-17-1 : Electronics > Video > Computer Monitors
el-17-2 : Electronics > Video > Projectors
el-17-2-1 : Electronics > Video > Projectors > Multimedia Projectors
el-17-2-2 : Electronics > Video > Projectors > Overhead Projectors
el-17-2-3 : Electronics > Video > Projectors > Slide Projectors
el-17-3 : Electronics > Video > Satellite & Cable TV

Categories el-17-2-1, el-17-2-2, and el-17-2-3 are leaf nodes—they have no children. Category el-17-2 (Projectors) is a non-leaf node because it has three children beneath it.

Why Taxonomies Exist

Standardized taxonomies serve several purposes:

• Site navigation: Customers browse products by category
• Search and filtering: Categories improve search results and filtering
• Platform requirements: Sales channels like Google Shopping or Shopify require products classified within their specific taxonomy
• Analytics: Categories help track product performance and customer behavior

Part 2: The Classification Challenge

Scale Problem

Modern product taxonomies are enormous:

• Google Product Taxonomy: Over 6,000 categories
• Shopify Standard Product Taxonomy: More than 10,000 categories

For stores with thousands of products, manually finding the correct category within these massive taxonomies becomes impractical. It requires deep taxonomy knowledge and is time-consuming, error-prone, and difficult to maintain consistently.

The Ambiguity Problem

Even with taxonomy knowledge, classification isn’t always straightforward. Consider these categories from the Google taxonomy:

267 - Electronics > Communications > Telephony > Mobile Phones
543513 - Electronics > Communications > Telephony > Mobile Phones > Contract Mobile Phones
543512 - Electronics > Communications > Telephony > Mobile Phones > Pre-paid Mobile Phones
543514 - Electronics > Communications > Telephony > Mobile Phones > Unlocked Mobile Phones

If you need to classify “Apple iPhone 16 Pro 512GB (Black Titanium)”, which category is correct? The product description doesn’t specify whether it’s contract, pre-paid, or unlocked. A human classifier faces the same ambiguity an automated system does.

Part 3: How Product Classifier Handles Taxonomies

Product Classifier uses AI to analyze product descriptions and automatically match them to the most appropriate category within your chosen taxonomy.

Basic Operation

1. Send a product description: Provide details about your product
2. AI analyzes the content: The system understands context, features, and product type
3. Receive the category: Get back the category ID, full path, and hierarchy

Handling Ambiguity: Non-Leaf Fallback

When the AI cannot confidently distinguish between leaf categories, it returns a more generic parent category instead. This is a feature, not a limitation.

In the iPhone example above, unless the product description explicitly states the phone type, Product Classifier would return:

267 - Electronics > Communications > Telephony > Mobile Phones

This generic parent category is accurate—the product is definitely a mobile phone—while avoiding an incorrect specific classification.

Leaf-Only Mode

Some businesses have rules requiring all products to be classified at the leaf level only. Product Classifier offers a leaf-only setting that constrains the AI to return only leaf categories. When enabled, the system will choose the most likely leaf node even in ambiguous cases.

Custom Instructions

You can guide the AI’s decision-making with custom instructions—natural language prompts that tell the system how to handle edge cases or apply business logic. For example:

• “Classify phones as unlocked by default unless description specifies contract or pre-paid”
• “Prioritize material over style for clothing items”
• “When product fits multiple categories, prefer the more specific one”

These instructions let you encode your business rules directly into the classification process.

Working with Multiple Taxonomies

Product Classifier supports:

• Shopify Standard Product Taxonomy: Optimized for Shopify stores
• Google Product Taxonomy: Essential for Google Shopping and advertising
• Custom taxonomies: Contact support to implement your own classification system

You choose the taxonomy for each classification request, allowing you to classify the same product into multiple taxonomies for different sales channels.

Getting Started

Understanding these taxonomy concepts helps you use Product Classifier effectively. Test classifications in the Playground to see how the AI handles your specific products, then integrate the API when you’re ready to automate classification at scale.