Multi-seller marketplaces and ecommerce stores use product category taxonomy structures to organize product groups so customers can find the products they're looking for in the least amount of clicks. The hierarchy structure makes the product catalog easier to search through refinement based on what the customer is searching for.
Google shopping categories are a list of categories outlined in the 2021 Google Product Taxonomy. There are 5,585 unique categories in the Taxonomy and are used by companies such as Google Shopping, Shopify, and many retailers as a baseline framework for defining product groups. The list of categories is commonly used as an internal structure for products or for the category ID required in Google shopping.
There are a number of key reasons why ecommerce brands use Google's product categories and the taxonomy structure to organize their products.
One of the most well known reasons for using Google's product taxonomy is to interact with primary shopping campaigns in Google Merchant Center. Google will automatically categorize your product data into a specific category based on title, description, price, brand, and GTIN, but does allow you to override the assigned category for some cases. Having your brand's products in the right google product category makes it easier for Google Shopping to align buyers to your Feeds.
Standardizing your products to google shopping categories increases the ease of Google ads campaign management and optimizing your keyword choice.
Multi-channel selling through Facebook and Instagram becomes easier to leverage as these platforms require the Google product category ID for categorization and sales tax reasons in direct checkout features.
Ecommerce platforms such as Shopify are starting to recommend the use of standardized taxonomies over custom categories. Shopify uses a variation of the Google product categories for what it called its standard product type.
Using the Google shopping categories to outline your on-site product catalog provides a concrete structure to your ecommerce store layout. This structure makes it easier to format product pages and catalog pages to rank in Google without cannibalizing your own pages. This also leads to keeping your product pages in line with the keywords you're ranking, making it easier for potential customers to navigate to the products they want to purchase. Customers being able to find products easily is a huge part of organic traffic conversions, as 79% of searchers will head to a competitor's site for the same product if they find the product difficult to navigate to.
Product data categorized into the GS1 global product catalog maps surprisingly well to the Google shopping categories with very few issues. GS1 categorized products can be easily mapped to the correct google product category with the official mapping file.
Deep learning models focused on ecommerce have allowed companies to automatically categorize products into the 5585 Google shopping product categories currently outlined. These models remove the need for constantly updating business rules or keyword matching logic that grows in complexity over time. Our automated product categorization model has learned a deep relationship between product data and the Google Product Taxonomy. Our models are well over 92% accurate and have been trained on millions of product records in the task of matching to a single predefined google product category.
Pumice.ai allows you to go from unstructured product data to clean, categorized, and high-quality data in minutes.
Pumice.ai's Google taxonomy categorization model requires just the product title and description for each record. Other popular attributes such as price, GTIN, and SKU can be included as well and become part of the description field.
CSV upload and direct API connection are the two ways you can start categorizing product data in Pumce.ai.
The key difference between the ways you can interact with the endpoints is the level of automation you can create. Direct API connection lets you easily integrate these ai models into your workflows for creating or updating product information. On the other hand, CSV upload lets you get started categorizing thousands of products in a bulk run in just minutes. Custom connections to ecommerce platforms such as Shopify can be built for you to further optimize your product information tasks.
We’ve integrated our fine-tuned GPT categorization model trained on millions of products into all of our plans. This model is well over 90% accurate and has been trained on products from Amazon, eBay, Macy’s and more. All you have to do is select the “google_product_taxonomy” from the dropdown menu and you can categorize your product data in seconds. Fine-tuned and enterprise users can adjust the outputs of this categorization pipeline to include information such as:
1. Confidence scores
2. “Best of” category results
3. Product information suggestion
Once you’ve started your task just sit back and relax! Your categorization data will be returned to you in the same format you passed it in based on the route needed (API vs CSV).
The generated product categories can be leveraged by passing this data to your PIM software or ecommerce platform. Your desired workflow is a key indicator to what you should do with this categorization data once generated and what the pipeline looks like. If you’re using the categorization endpoint as a piece of an automated workflow such as product creation then building downstream systems to automatically integrate product information into output systems is key. Automation to integrate the product data improves ROI and reduces manual effort required for product information management tasks. We build custom integrations for upstream and downstream tasks to systems such as:
- Ecommerce Platforms (Shopify, Woocommerce)
- Zapier
- PIM Software
- Make
- Internal product databases
- Google Shopping Feed
The image above shows an architecture diagram of a fully automated product information management pipeline using Pumice.ai endpoints.
Pumice.ai is an AI leveraging PIM enhancement platform built to help ecommerce stores and multi-seller marketplaces automate product and data information tasks at scale. Take the manual focused nature of PIM software to the next level with our baseline models proven to reduce costs and time. Have a custom use case or unique product information setup? Fine-tune these same models specific to your use case to see huge increases in accuracy. Contact us today to schedule a demo and learn more about how you can start using ai to automate your PIM tasks today.
Here’s a list of the Google Product Categories as provided by Google.
