Be flexible to change and expand
Posted: Wed Jan 22, 2025 4:09 am
It is unlikely that a data model defined at the beginning will continue to be used without changes in the future. Therefore, when defining a data model, it is important to consider some "space" that anticipates future possibilities. For example, should product classifications be limited to just two types, "standard products" and "custom products," and will this increase in the future? Can drawings be linked not only to products but also to series? This means considering definitions that anticipate such things.
However, while allowing flexibility to the data model makes it easier to change and expand in the future, it also increases the complexity of the initial data model why you need this database considerations. It is difficult to get the balance right when it comes to flexibility, as being too generalized or too specialized is not good, but PIM products allow you to add, update, or delete attributes and relationships later, and some have mechanisms for partial updates and bulk replacement, so it is a good idea to check the functionality in that area.
Check the unit of connection with external systems
The general integration pattern for PIM is to import master data, process it in PIM, and then integrate it with websites or other systems. If you do not check the number and unit of data to be integrated at one time, it will take time to integrate the differences and it will be difficult to process the data in the integrated system. Integration with external systems is a case-by-case matter, so it is difficult to suggest a general rule. However, by considering the frequency of integration and the number of data items, and deciding on the optimal integration unit, you can reflect that in the data model. For example, promotion information that is frequently updated in EC can be stored as a separate data model that has a 1:n relationship between products and promotion information, and only updates to the promotion information are reflected.
This depends on the structure of the receiving system, so it is not possible to achieve this with all systems, but by streamlining the connection between systems, it is possible to increase the frequency of information updates, which has the benefit of always providing fresh information to customers.
However, while allowing flexibility to the data model makes it easier to change and expand in the future, it also increases the complexity of the initial data model why you need this database considerations. It is difficult to get the balance right when it comes to flexibility, as being too generalized or too specialized is not good, but PIM products allow you to add, update, or delete attributes and relationships later, and some have mechanisms for partial updates and bulk replacement, so it is a good idea to check the functionality in that area.
Check the unit of connection with external systems
The general integration pattern for PIM is to import master data, process it in PIM, and then integrate it with websites or other systems. If you do not check the number and unit of data to be integrated at one time, it will take time to integrate the differences and it will be difficult to process the data in the integrated system. Integration with external systems is a case-by-case matter, so it is difficult to suggest a general rule. However, by considering the frequency of integration and the number of data items, and deciding on the optimal integration unit, you can reflect that in the data model. For example, promotion information that is frequently updated in EC can be stored as a separate data model that has a 1:n relationship between products and promotion information, and only updates to the promotion information are reflected.
This depends on the structure of the receiving system, so it is not possible to achieve this with all systems, but by streamlining the connection between systems, it is possible to increase the frequency of information updates, which has the benefit of always providing fresh information to customers.