Use Case 01

Automatic tagging in online catalogues

Realized high-speed automation of product registration (10 million items per day) which led to a 100-fold increase in merchandise in company EC (from 1 million to 100 million).
Automatically extracts merchandise attributes from suppliers’ unique product photos and text data, then categorizes them into the “company’s unique product categories.”

Improvement in sales
Improvement in productivity
Securing of labor force
Cost reduction
Image analysis
CrowdANALYTIX

Business challenge

Retail businesses with physical stores made it their strategies to expand purchase opportunities through their own EC with the rise of EC platformers.

It was essential to realize speedy product registration and an increase in product related information in order to expand the number of product lines in their own EC.

Specifically, it was then necessary to have a

system that extracts product categories and attributes new products quickly based on the data of various products provided by multiple suppliers, in a way that fits their own product databases.

By realizing the above, you can propose products that coincide with the needs of each customer in your own EC in a timely manner, in addition to targeting the further expansion of purchase opportunities.

With traditional e-commerce site, because it was data provided by suppliers on a wide variety of products, the data was not utilized as the company’s own data in a uniform manner.

Because categorization and registration into product categories of their own companies was done manually, efficiency and accuracy lacked in uniformity depending on the person. Moreover, since product registration was time consuming, there was always a situation in which people had to wait for new products to be registered, making time-to-market a big challenge.

In terms of costs, costs went up in proportion with the number of registered products. Companies were having business challenges in which they were unable to use a strategy of increasing product lines.

In response to these challenges, we have created the company’s own database which automatically determines product categories and merchandise attributes through AI from product images and text data, no matter the supplier or product. Through this, the time until product registration was significantly reduced, product lines were significantly increased, and time-to-market of new products was now done much faster. Moreover, uniformity of product categorization and attribute extraction was improved, a system was made to sustainably improve accuracy, and each customer’s needs are now reflected in a timely manner.

※Scroll to view.

Before

It was difficult to attach the company’s unique product categories and attributes as there is a wide variety of product lines from the supplier.
The categorization and registration of products was performed by humans, making it time consuming and incapable of being carried out for the huge amount of product lines.

After

AI automatically determines the product category and merchandise attributes, constructing a unique database for the company.
Manual labor was automated and time-to-market of new products was reduced