How marketplaces use the power of AI to migrate their data from old taxonomy to new

Taxonomy development has been ad hoc across all domains, not just fashion. With fashion, it gets that much more complex as there is no universally accepted nomenclature for describing products. Some call it long-sleeves and some call it full sleeves, who is to say which is better. So, brands have typically adopted whatever generic terms they have been using within their companies to build their first versions of their taxonomies. But as they evolve digitally, enterprises progressively realize the value and feel the need to add more dimensions and characteristics to their taxonomies.    

But fashion retailers typically hesitate to migrate to a new taxonomy even if it is better, as the effort of migrating all the data is complex and time-consuming. But there comes a time when it becomes imperative to change and a catalog manager's nightmare comes true. If you have done cataloging on excel sheets with 100’s of columns you will understand the complexity of the issue on hand.

‍

Let’s take a typical example that most of our customers are experiencing when they are migrating from one taxonomy to another. Adding new attributes or values is relatively easy. But most of the time, important changes are made in the information architecture such as changing the attributes or categories which form the column names on lists and libraries. This makes the migration a little more difficult since the column names no longer match the source and destination. Sometimes, it’s even worse when column types are no longer the same and things are re-defined or hierarchies are changed.

Streamoid helps brands and marketplaces migrate their data from one taxonomy to another using Catalogix their auto-tagging and cataloging solution.  

How it works is looks simple, but it was very complex to build. Streamoid has one of the largest and most well-researched fashion taxonomy in the world. We use computer vision to identify and tag all the relevant attributes in the garment based on our taxonomy. We use NLP to check if there are any different values that are then flagged for a manual QC. Once data accuracy in Streamoid taxonomy is verified, it gets populated into the new taxonomy sheets at the click of a button.  

Catalog managers can now relax and not shudder at the thought of a taxonomy migration. All they have to do is hand over the old and new templates to us along with the data feed and wait for a clean set of data to be handed back to them.

Good data is critical to not just understanding what is going on but to identify new opportunities as well. With IoT gaining popularity it is going to be more and more critical. Now with the taxonomy migration piece solved you need not hesitate to improve your taxonomy and drive your business forward.

‍