Crystallize

Information Architecture for multi channel PIM

The heart of any ecommerce solution are the products that are offered. You need to curate rich products descriptions to sell and it all needs to be stored in a way that is reusable across multiple channels and in a format that developers can use regardless of language and channel. A proper information architecture for product information is key to avoid mess over time. Product information management is the core.

Products in all shapes and sizes

The product information management in Crystallize can manage any type of products from one-off sales to recurring subscriptions. Products can either physical goods to be shipped or virtual products like enabling a software service.

One size does not fit all. Products come in all shapes and sizes. For that reason you need to organize the structures or groups of products by the different properties they describe. We call this shapes in Crystallize.

If you consider a t-shirt and a pair of jeans they might be fine sharing the same describing shape. What if you in the same shop are selling access to a streaming service? Would you use the same properties to describe a t-shirt and a monthly subscription to a streaming service?

This is where shapes and components come into the picture. Simply put a shape defines a set of components that make out the way you can describe and enrich a product. Remember we want both to have a good structure and we want rich information to sell the products.

Semantic Structures for PIM shapes

Ideally we bring semantics into the description of different products. If we have a good semantic structure we can for example compare products of similar shapes. Like all clothes made of cotton in Italy.

A simple example of a clothing shape:

  • Name
  • Short description
  • Price
  • Product picture
  • Properties:
  • Fabric
  • Manufacturing country
  • Eco certification

Here we define a properties component that is common for all clothing products. You typically want to know if it is wool, cotton. If it is manufactured in Europe or somewhere else. Does the piece of clothing have some sort of eco certification? A describing shape like this works great for clothes, but what if we have a subscription to a service?

A digital service shape would be different:

  • Name
  • Service description
  • Monthly price
  • Picture
  • Teaser video
  • Terms of service

These are simple shapes, but still show that you should ideally have a different shape to describe clothing than you would use to describe a digital subscription service.

GraphQL to query product information fast

Assuming you have a neatly structured set of shapes describing your products. That product information is curated properly. And finally that the product information is stored and are available in a properly structured format like e.g. GraphQL. This is the approach we have in Crystallize.

With GraphQL developers can easily retrieve structured rich product information from the Crystallize repository - in real time.